The first local investor I ever pitched looked like a shark.

I do not mean that metaphorically. I mean the man had small, flat eyes set too wide apart, a mouth that turned down at the corners even when he was being polite, and he sat behind his desk in a Rampura office that smelled of cigarettes and stale air-conditioning the way some Dhaka offices have smelled since the 1990s. I had been introduced through a chain of relatives. Aunt’s husband knew his cousin. The cousin had said this uncle “did investments.” I had imagined an office with a glass partition and a coffee machine. What I got was a man in a half-sleeve shirt who started the meeting by asking me how much money my father made, and ended the meeting by offering me five lakh taka for forty-nine percent of the company.

He was not even shy about it. He smiled while he said the number. I remember walking down the stairs of that building afterward and standing on the footpath outside, the heat coming up through my shoes, a rickshaw bell ringing somewhere behind me, and feeling the specific kind of humiliation that comes from realizing the person who just tried to rob you also thought he was doing you a favor.

That was investor number one. There were many after him.

Every Door In Dhaka, Then Every Door On The Internet

I wrote the same email a thousand times in 2015 and 2016. Two paragraphs, a link to a deck, a line at the bottom asking for fifteen minutes. I sent it to every plausible name in Bangladesh I could find. The wealthy families who occasionally wrote checks into businesses. Bank executives who would take a meeting out of politeness. The textile money. The real estate money. The cousin of the cousin who “did investments.”

Most did not reply. Some did. A handful turned into meetings, and the meetings were worse than the silence. An hour in a conference room with a man twice my age who would nod through the pitch, ask one or two questions to be courteous, then explain to me, gently, with sugar in his tea, that money in Bangladesh worked through land and factories and contracts with the government, and not through whatever I had just described One of them took me to lunch at a Gulshan club, fed me biryani, and ended the meal by advising me to find a real job. He was not being cruel. He was being honest about the world he had lived in for thirty years. The world he had lived in for thirty years did not have a slot for the company I was trying to build.

When Bangladesh ran out, I went to the internet. Y Combinator. Rejected. 500 Startups. Rejected. Techstars. Rejected. Seedstars made me a finalist and then cut me on the final round. I applied to every program on every continent with an open form and a half-credible website. Every single international accelerator said no. The rejections came in at three in the morning because of the time zones, and there is a specific loneliness in reading a Y Combinator rejection at four AM on a mattress on the floor of a one-room apartment in Bashundhara, with the dawn azan starting up somewhere outside the window and the sound of the chacha at the tong setting up his stove on the street below.

Airbnb went through the same thing in 2008. Brian Chesky kept the emails. Seven Silicon Valley investors, every one of them credible, all of them passing on what is now a company worth more than the largest hotel chains combined. The investors were not stupid. They were pattern-matching against companies that had worked before, and Airbnb did not match. Neither did we.

The Pitch Is Not A Static Object

Here is what I did not understand for the first year: the pitch is not a thing you write once and read out in every room. The pitch is a living document, and every no is information you are supposed to fold back into the next version.

I started writing down the questions I could not answer. The first month I had ten of them. The second month I had twenty. By the sixth month I had a list of forty questions investors had asked me that had stopped me cold, and I had answers for every single one, because every time a question caught me out I went home and worked on it until it would not catch me out again. My pitch in month twelve had almost no words in common with my pitch in month one. The company had not changed. The pitch had been rebuilt forty times by the people who said no.

The other tactical thing I learned, the one I wish someone had told me on day one, is this: do not start your pitch with the investor you most want to win. Burn it on the investors you do not. Go to the small fish first. The regional accelerators with bad reputations. The angel groups that meet in coffee shops. That rich uncle. Pitch to them. Take the rejection. Take the questions. Take the friction. Use them as sounding boards, because the polished version of your pitch only exists on the other side of fifty bad ones, and you cannot reach the polished version any other way. By the time you walk into the room with the investor you actually want, you will have answered every question in that room a hundred times in rooms that did not matter.

The Grind Does Not End When You Get Big

By 2019, Pathao was the largest tech company in Bangladesh. Two million users, fifty thousand drivers, profitability on a contribution-margin basis, national coverage, household name. By every metric the 2015 investors had said was impossible, we had become the company they had told me to stop trying to build.

And in 2019, when we needed the next round, they said no again.

The pitch was different now. US lawyer. Audited financials. A deck designed by someone who understood typography. Warm intros to every partner at every relevant fund in Asia. None of it mattered. The post-Uber-IPO freeze had killed the appetite for ride-sharing exposure across the region. We pitched Naspers, the South African giant that owned a piece of Tencent. Passed. Tiger Global. Passed. SoftBank when the Vision Fund still had a pulse. Passed. Sequoia India, Lightspeed, funds in Hong Kong and Bangkok and Dubai whose names I have honestly forgotten.

Spotify went through the same thing in slow motion. Daniel Ek spent thirty months between 2007 and 2009 pitching record labels on streaming. Every meeting ended the same way, polite interest, no license. He kept going because outlasting the industry’s resistance was the only move available to him. When Sony finally signed, the others followed within months. The thirty months were not a failure of the pitch. They were the work.

Pipeline, Not Prayer

The mistake many first-time founder makes is treating one investor like the whole company depends on them. You meet someone, you put your whole life into that conversation, you wait two weeks for the email, the email is a no, you collapse. At that pace you run out of money before the first row of your list is finished.

The fix is a pipeline. Hundred names at the top. Thirty take a meeting. Twelve become real conversations. Five go into diligence. Two send a term sheet. One closes. What a pipeline gives you is parallelism. Investor A is in diligence, Investor B is in second meeting, Investor C just took the warm intro, Investor D got prospected this morning. When Investor A drops out, which Investor A will, Investor B has already moved up. No single conversation can kill you.

This is exactly how a good sales team runs against enterprise customers. Same stages, same conversion math, same arithmetic. Founders who came up through sales raise capital faster than founders who came up through engineering, because they have been running pipelines their whole career and the rest of us have to learn it from scratch.

The Only Skill That Matters

Everything I have told you about fundraising is also true about the rest of your life.

It is true of hiring. The best person you ever hired turned you down twice before they said yes, and the next best person you ever hired was someone you almost rejected because they failed the first interview. It is true of sales. The customer who became your biggest account ignored your first nine emails. It is true of dating. The relationship that became your marriage almost did not happen because one of you was about to give up the night before the second date. It is true of friendship. Every close friend you have was, at some point in the early conversation, someone you had to keep showing up for past the moment they could have stayed a stranger.

Every form of success on earth runs on the same engine, which is the willingness to be rejected and to keep going anyway. The people who succeed are not the people with more talent or more luck or richer parents. They are the people who can sit across from a shark in Rampura, walk out humiliated, and email the next investor on the list that same evening. They are the people who can read a Y Combinator rejection at four AM and apply to 500 Startups before they sleep. They are the people who can pitch the same numbers in the seventy-third room with the same energy they brought to the first.

If you cannot do this, the world will sort you into the version of your life that did not require it. Most people accept that sorting. There is no shame in it. There is also no escape from it, because the version of your life that does not require rejection is also the version that does not require courage, and the things that change anything for anyone, your company, your love, your friendships, your sense of who you are, all sit on the far side of someone telling you no.

The yes is the receipt. The work is the grind. And the grind, if you let it, is the only thing that has ever separated the people who get the life they want from the people who get the life that is left.

I had to dig up an old photograph for this post. An under-construction room in an under-construction building, all of us crammed in for the frame. We’re wearing our first branded Pathao shirts, red, ordered from the cheapest vendor we could find, the kind of red that bleeds in the wash. You can see in our faces that we hadn’t slept properly in weeks. I’d like to believe that you can also see our optimism. Nobody in that photograph knew what they were doing. But everyone in it had decided, in their own way, that they were going to figure it out.

I didn’t take a lot of photographs from those days. We were too busy working. But now I wish I did. Back then I thought Pathao was a logistics company. I thought the deck in my laptop bag, the one with the market sizing and the five-year projection, was the company. It took me three years and roughly twenty bad hires to figure out I had it backwards. The thing you build is downstream of the people you build it with. Get the people wrong and no plan survives contact with reality. Get the people right and they’ll rewrite the plan three times before lunch and the company will still be standing at dinner.

The Territory Manager from the Tobacco Firm

In 2017, when we were hiring twenty people a week, I made the most expensive hire of my life up to that point. A territory manager from one of the big multinational tobacco firms. Massive resume. Sharp shirts. Spoke in case studies. He got paid more than I did. I wasn’t even mad about it, I told myself that was the whole point. We desperately needed someone with more experience than any of us had. I was twenty-six. My head of operations was twenty-three. We had no idea what we didn’t know. This guy had run distribution networks across half of Bangladesh for a company that had been selling things in this country for longer than I had been alive. On paper, this was a steal at any price.

He lasted four months.

First week, he scheduled an all day meeting about vision and mission. Slide decks were prepared in advance. While we brainstormed brand pillars, our drivers in Mohammadpur were being ticketed by traffic police. Second week, an engineer messaged me asking whether the new hire was now his boss, because he’d been told to “loop him in” on every product decision. Third week, he was calling our riders “the field workforce”.

He wasn’t a bad person. He was an excellent operator. His default mode was to make a plan, present a plan, defend a plan, execute a plan. Ours was to do something, see what broke, fix it, do something else. Two different operating systems. Same machine. They didn’t run.

The cost wasn’t his salary, though that hurt. It was the trust the team lost in my judgment, because every one of them saw by week two what I refused to see until week sixteen. Netflix figured this out earlier than most. Reed Hastings introduced the keeper test. For every person on your team, ask whether you’d fight to keep them if they walked in tomorrow with another offer. If the answer is no, you already have your answer. The wrong hire isn’t dead weight. It’s a magnet. It pulls the company in a direction you didn’t choose. I should have fired him in week two. I fired him in week sixteen.

The lesson I missed was simpler than I wanted it to be. I hadn’t hired the wrong person. I had hired the right person for the wrong company. He would have crushed it at a sixty-year-old enterprise with stable distribution and quarterly planning cycles. He couldn’t function at a place where the answer to “what’s the plan” was “we’ll know by Thursday.” Pedigree is a signal about the system someone was trained in. It tells you almost nothing about whether they can work in your system.

Attitude Over Resume

Our first HR manager taught me the opposite lesson. I told him, day one, his job was to find me 100 riders. He blinked. Asked from where. I admitted I had no idea. He spent his time sifting through thousands of badly written CVs in bdjobs.com and bikroy.com, and having 20 meetings a day with blue collar workers checking if they are not creeps.

He treated “I don’t know how” as a starting condition, not a stopping condition. That’s the only variable that ever mattered. Two sentences: “I don’t know how to do that” vs “I don’t know how to do that yet, let me figure it out.” They look identical. They are completely different operating systems. The first ends a conversation. The second starts work.

The MIT-educated engineer who quit two months in when our staging environment kept crashing? Wall. The girl from a Bangla-medium school in Mirpur who taught herself SQL on YouTube in three weeks and went on to run our growth analytics? Information. The smartest people I’ve ever met failed at Pathao. The hungriest stayed and ran the company.

This matters more in Bangladesh than San Francisco for a reason. The default career advice from parents to children in Bangladesh is to find a a stable government job and never leave it. The whole society is wired to optimize for security. Which means the resume class, the pedigreed ones, are precisely the people trained from birth to flinch. They execute known plans beautifully. They collapse the moment the plan dies. And in a startup, the plan always dies.

Airbnb’s first hires were the inverse. Chesky famously asked candidates whether, if they had one year to live, they’d still take the job. Most said no. The yeses became Airbnb. Same filter. People who don’t need permission to be all-in.

Hire Slowly. Fire Fast.

Almost every founder says these words. Almost none of us practice them. Including me.

When you’re growing 100% month over month, slow hiring feels like a death sentence. We told ourselves we couldn’t afford to be picky. We were wrong. The cost of a slow hire is one missed week. The cost of a wrong hire, factoring the time you spend managing them, cleaning up after them, firing them, and the morale damage to everyone around them, is closer to a year. Slow hiring is the cheap option. It just feels expensive in the moment.

The reverse is also true. The moment you know someone is wrong, you already know. You don’t need another quarter. You don’t need another review cycle. You know in your gut, in the way you tense before your one-on-ones with them, in the way you find yourself doing their job. The reason we don’t act isn’t uncertainty. It’s discomfort. So we wait. While we wait, the best people on the team watch us wait, and lose faith.

The Bangladeshi cultural pressure on this is real. We’re a relational society. You don’t fire someone whose family knows your family, whose father called your father to thank you for the job. The whole social contract is wired against the clean cut. Founders here drift into the pattern of carrying the wrong people for years, paying with the morale of the entire company for the comfort of one conversation we don’t want to have. I did this. I’m not proud of it.

What eventually worked was crude. Every six weeks, the leadership team independently ranked our reports on two axes: contribution to outcomes, and contribution to culture. We’d compare lists. Anyone in the bottom of two or more was a candidate for a hard conversation. The point wasn’t the ranking. It was the calendar. Without a calendar, the conversation never happens.

What Right Looks Like

When we eventually shut down our customer support center in Rampura, the team leader who ran that floor asked one thing. He wanted to tell his own team before I did. I said no, it had to come from me, I owed them that. He said, with respect, I was wrong, because they wouldn’t believe it from me, and they needed to be able to ask questions only he could answer. He was right. I let him do it.

What he did next is the part I won’t forget. After the layoff, he stayed. Spent three weeks, on his own time, helping every one of his people find their next job. Called his contacts. Wrote recommendations. Coached them through interviews. Ninety people. He’d just been laid off himself. He didn’t have to do any of it. He did, because in his head, his job wasn’t over until his people were okay.

You can’t teach that. You can’t interview for it. You can only spot it after the fact, and build a company that produces more of it by hiring other people like that and protecting them when the machine tries to grind them down.

Amazon’s Day 1 doctrine isn’t really about being a startup forever. It’s about being run by people who behave like founders. The thing that compounds is the quality of the humans. Everything else is downstream.

I started Pathao thinking I was building a logistics company. I ended Pathao knowing I’d been building a team the whole time. The logistics was the byproduct. The team was the product. If I’d understood that on day one, I’d have spent half the time I spent on the deck and twice the time in the recruiting room. The bus is going somewhere. The road will turn ten more times before you get there. The map you brought will turn out to be wrong. What you have when you arrive is the people who got on the bus with you. Get the people right and you’ll figure out the road. Get the people wrong and the road won’t matter.

Imran bhai had been with us since the second hire. He had carried boxes up the unfinished stairwell in Chairmanbari and slept on a desk the night our payment gateway crashed. People loved him.

By 2019, we had crossed five hundred, and Imran bhai was drowning. Vendor invoices paid twice. A campaign shipped without legal sign-off. A junior came to me directly because she could not get a decision out of him for two weeks.

“Bhai, ektu time lagbe,” he said in our last review. I gave him another six months.

What I refused to see was the quieter cost. New hires asking what he did all day. Team leads who knew his mistakes never landed on him. Interns noticing the founder’s favorite got a free pass.

Understanding

In Bangladesh, a job is rarely just a job. It is the household. School fees, rent, medicine for a parent, the only thing between dignity and disaster. There is no social security here. No unemployment cheque after a layoff. People live month to month, and losing a job in Dhaka has nothing in common with losing one in Berlin.

Other cultures have built their own compacts around this. Toyota and a generation of Japanese and Korean companies anchored themselves to lifetime employment, where letting someone go was nearly unthinkable. Every country writes its own deal with work. Ours has been forged by precarity.

Our work culture orbits around a single heavy word. Understanding. “Bujhen amar obostha.” Employees ask for it. Employers extend it.

Understanding is a beautiful thing in a neighborhood. It becomes a slow poison in a scaling company. Every month of understanding extended to one underperformer is paid for by the four people around them.

The Keeper Test

Reed Hastings built Netflix around a single brutal question. If this person walked in tomorrow and resigned, how hard would you fight to keep them? If the honest answer is “I would let them walk,” they should already be off the team.

A second move follows. Every cycle, cut your bottom performers. Especially when the team is strong. The only way to raise the floor is to keep raising it.

The Two-Month Notice Trap

Bangladeshi labour law gives you two months’ notice. At Pathao, we set our policy at two months on purpose, thinking we were being responsible.

It was theatre. I never saw a handover that genuinely needed more than two weeks. What happens in those weeks is predictable. The person takes interviews on company time. Calls go unreturned. Slack messages get the slow-drip treatment. Their heart left the building the day they got the letter. Pay the notice. Send them home that afternoon.

Who You Actually Owe

When you decide to let someone go, the question is who you are actually protecting. The Bangladeshi instinct sends every drop of empathy toward the person being fired, which is the easiest choice.

What you owe is the company. The co-founders who bet years on the same thing you did. The investors who wrote checks on your promises. The three hundred others on payroll watching whether you have the spine to protect what they are building.

How to Fire

Step one is the performance appraisal. Sit down, name what is working, name what is missing. If HR is in the building, write it down so the record exists on paper.

Then give them three months. Long enough to turn it around if the will is there. Inside that window they move or they don’t.

If they don’t, the second meeting goes on the calendar. Twelve minutes. You tell them the role and the person are no longer a fit, you outline the severance, you give them room to respond. You make it about the role, never about who they are. Both of you already know what is coming.

This whole sequence has two purposes. A paper trail that answers for itself later. And dignity. None of it should feel personal. The role asked for something the person could not deliver, three honest months tested that, and a clean decision followed.

Pay the severance generously. Your remaining team is watching whether you treat tenure like humanity.

I finally let Imran bhai go that December. He had known for weeks. He found a new role inside two months. He still messages on Eid. The team he led became one of our strongest within the quarter. I had waited six months too long, and every month cost something I could not get back.

Be less understanding. Be more responsible. Those are two different jobs, and confusing them is how good companies die slowly.

My software stack evolves frequently as I experiment with new apps. Here’s my current setup:

Utter (Free) ✨✨✨

AI Powered voice-to-text that I run locally with NVidia Parakeet + Apple Intelligence for cleanup. No lag, since it’s run all locally. Completely changed how I use my computer and write. In fact, I’m writing this using Utter right now :)

Raycast (Free) ✨✨

A collection of powerful productivity tools all within an extendable launcher. Replaced Spotlight + PastePal from before. Do take some time to configure this: I have custom shortcuts for snippets copy and paste search Spotify and more.

Things ($49/forever) ✨

Honestly, Apple Reminders does a great job at a to-do list. However, I’m a sucker for a clean user interface and extensibility that Things offers. I feel like I can design my life around my to-do list instead of the other way around.

Markedit (Free) ✨

MarkEdit is a free Markdown editor, for macOS. It’s just like TextEdit on Mac but dedicated to Markdown. Since I now need to edit Markdown files a lot more to write skills for Claude, a dedicated Markdown app has become necessary

Granola (Free) ✨✨

AI powered note-taking tool that works amazing. I was using it a lot more when it was completely free but now it costs $14/month. I am exploring local alternatives which do exist but nothing that works as good as Granola yet

Switch (Free) ✨✨✨

⌘‑Tab that acts like Alt-Tab in Windows. Minor but great quality of life improvement

Castly (Free) ✨✨

Loom alternative that is free and local. I use it to give quick feedbacks to my team while sharing screen

What I continue using from my 2025 Stack:

Superhuman ($30/mo) ✨✨✨

Superhuman is my go-to email client. It's fast… really fast. Clean interface, intuitive shortcuts, built in AI features and email tracking covers everything I need to go through hundreds of emails every day. Going through all my emails to inbox zero takes a quarter of the usual time, and makes working fun.

Worth it if you are living out of your inbox like me.

Alternatives: Simplehuman (free), Simplify Gmail ($3/mo)

Dashlane ($3.30/mo) ✨✨✨

Never use the same password in more than one site. Dashlane generates strong passwords, autofills credentials and OTPs across all devices, and stores Passkeys (I never want to get ecosystem locked by giving Apple all my password/passkeys).

100% worth the cost of a small coffee per month.

Note: always turn off SMS 2FA in every site. Sim swap attacks are surprisingly accessible.

NordVPN ($90 for 2 years) ✨✨

How one can live in Bangladesh without VPN is beyond me. Unlock accesses to websites and remove yourself from prying eyes. Always on VPN.

Alternatives: Surfshark ($60 for 2 years)

PDF Expert ($140, one-time payment) ✨

Yes, I am one of those guys who pay for PDF Softwares. But if you have to deal with as many PDFs as I do - PDF Expert is excellent. I like editing any PDF files easily (which reduces going back to Docs to make edits and exporting again…) and thus save time.

zCal (Free) ✨✨✨

Free calendly alternative that can support multiple links and multiple calendar accounts. Notion calendar still doesn’t have reusable scheduling links.

Notion Calendar (formerly Cron) (Free) ✨✨✨

Finally, for scheduling, I rely on Cron (Notion Calendar is such a bad name). Again - keyboard shortcuts, speed, simple workflows make it the superhuman for calendars. The time travel feature is such a killer for scheduling meetings before I travel, and the join from menu bar widget is so very helpful.

Cloudflare (Free) ✨✨✨

I manage all my personal websites through Cloudflare. Free unlimited email forwarding + workers help convert any Notion page to a website (makes easy to use landing pages like istomorrowhartal.com). Super simple and fast to use DNS management!

Franz (Free) ✨✨

Managing WhatsApp, Telegram, Slack, Discord, Messenger etc etc as separate apps takes a lot of mental toll. Too many messaging apps! This consolidates all of them into one window.

Hevy (Free or $80/lifetime) ✨✨✨

Workout tracker and planner. Easy to make routines and track progress.

Related, I get my routines from BuiltWithScience

Airalo ✨

e-Sim for iPhone so I can have cheap data on demand wherever I travel. Essential and easy to use for frequent travellers. The pricing of these eSims turn out to be the same or cheaper as you would buy from the airport (I checked).

Referral code that get’s you $3: HUSSAI6543

Daylio ($20/year) ✨✨✨

Been using this for the last 5 years to act as my mini-diary and mood tracker. It’s my minimum viable journal. Only tool I have been paying for 5 consecutive years.

Rabby Wallet (Free) ✨✨✨

The best wallet for all EVM chains. Smooth multi-chain experience and improved secruity as they do a pre-sign check. Makes it easy to manage multiple wallets as well.

Referral code: ELIUS

Alternative: Zerion Wallet for iOS/Android

Headspace ($70/annually)

My go-to meditation app. Andy Puddicombe’s voice always calms me down.

Here’s a hack: if you email their support@, they will more often than not give you a 40% discount. Just ask :)

What has been removed from my 2025 Stack:

• Arc was replaced by Google Chrome, as Arc has been discontinued and Google Chrome now has a sidebar for Tabs. Not exactly the same things, but comes close.

• PastePal/Paste has been replaced by Raycast.

• SideDrive, Height.app Figma has been removed because I don’t use it enough.

Fahim Saleh did not find me through a LinkedIn post or a pitch competition. He found me through a joke.

I had built a website called IsTomorrowHartal? - plain white background, Arial font, a yes or a no. That was it. During a period when Dhaka was paralyzed by political strikes and everyone was burning the same question into every WhatsApp group and Facebook thread, I just answered it in one click. The site got 40,000 views on day one. 80,000 the next.

Fahim saw it floating around and messaged me. “How much to sponsor this?” I threw out $200 expecting a laugh. He replied in seconds: “Cool. Send me payment details.”

A few days later I was sitting in his office at HackHouse in Banani . Wooden floors, glass walls, whiteboards full of actual ideas. He was building a venture studio. The plan was simple: take business models that already worked in New York or Jakarta and make them work in Dhaka. He had the vision, the money, the track record. What he didn’t have was someone technical who actually lived here, understood this city, knew how the roads worked and which government office you needed to navigate and what a Bangladeshi user would and wouldn’t do on a smartphone.

That was the gap. And I fit it exactly.

He needed me because I could build the product, and because I was from here. I understood the chaos. I could navigate it. That’s the only reason I became his co-founder at HackHouse. Not because we were old friends. Not because I impressed him at a networking event. Because I filled a specific, critical hole in what he was trying to build.

That’s the first thing most founders get wrong. They don’t go looking for what they’re missing. They go looking for someone they already like.

Know Your Gaps Before You Go Looking

Before you even start thinking about who, you need to be brutally honest about what. What does this company actually need to survive? Not what would be nice to have. What is it that you literally cannot build without?

Every company needs at least two roles: a builder, and a seller.

Fahim’s answer was obvious: local technical execution. He was Bangladeshi by heritage but a New Yorker by everything else. He had the capital, the frameworks, the Silicon Valley thinking. But he couldn’t build the product, and he didn’t know how to move inside Bangladesh. Without someone who could do both, HackHouse was just a well-funded idea.

My answer, when I later co-founded Pathao, was different. I needed someone who could run ground operations while I stayed close to product. Someone who could negotiate with a government office in the morning and manage fifty riders in the afternoon without losing their composure at either.

The gap tells you who you need. Start there. Don’t start from the person and then work backwards trying to justify the choice.

Build Shared History Before You Build a Company

Fahim and I had months of working together at HackHouse before anything became official. We saw how each other thought. We disagreed. We moved fast. We had friction and got past it. By the time real pressure came, we weren’t strangers still figuring each other out. We already had a language.

This matters more than people admit. A co-founder relationship you jump into cold is the startup equivalent of marrying someone you met at an airport. It might work out. But you’re gambling the most important relationship in your company’s life on zero data.

Work on something together first, even something small. A contract project, a shared problem, a prototype built over a weekend. Watch how they handle setbacks. Watch what they do when you disagree. Watch whether they go quiet or go loud when things get hard. That’s your real interview. Everything else is just conversation.

This Is Closer to Marriage Than You Think

People say choosing a co-founder is like choosing a spouse. I used to think that was a metaphor. Now I think it’s an understatement.

When Pathao was in full build mode, I was at the office or on calls for ten, eleven, twelve hours a day. Some days more. My co-founders were the first people I spoke to in the morning and the last I messaged at night. We ate together, panicked together, argued about cash flow at 1am together. There were stretches where I saw my co-founders more than I saw anyone else in my life, more than family, more than friends, more than the person I was supposed to be dating.

Think about what that means before you sign anything.

In a marriage, you choose someone you want to build a life with. In a co-founder relationship, you’re choosing someone you’re going to solve problems with for ten to twelve hours every single day, under financial pressure, with other people’s livelihoods at stake, with investors watching, with a market that doesn’t care about your feelings. That is a more intense operating environment than most marriages ever face.

And just like a marriage, the cracks don’t show up on day one. They show up in year two when you’re exhausted and the money is tight and you realize you each had a completely different picture of where this was going. The resentment builds quietly. The small irritations compound. One person thinks they’re carrying more than their share. The other thinks the same. And suddenly you’re not building a company together, you’re managing a relationship that’s slowly breaking.

I’m not saying don’t do it. I’m saying treat it with the same seriousness. You wouldn’t propose to someone you met twice. Don’t hand out equity to one either.

Align on the Personal Stuff First

Before the product, before the strategy, before the deck, sit down and have the uncomfortable conversation about what each of you actually wants.

One person wants to build something big and sell it in five years. The other wants to build something that lasts and hold it forever. One person has three kids and needs to be home by seven. The other has no obligations and will work until they physically collapse. One person needs a large exit to justify the sacrifice. The other would be happy with a profitable small company.

These mismatches don’t explode immediately. They ferment. And they always surface at the worst possible moment, usually when the company is either close to success or close to collapse, the two times you can least afford a crisis at the top.

Get it all on the table before you start. You don’t need identical goals. You need compatible ones, and you need to actually understand what the other person is walking in with.

Two to Three. No More.

Keep the founding team small. Two to three people, four at an absolute stretch.

I have seen nine-person founding teams walk into a room. That is not a founding team. That is a committee with equity. The more co-founders you have, the more complex every decision becomes, the harder the equity conversation gets, and the easier it is for misalignment to quietly hollow out the structure from the inside while everyone is still smiling at each other in meetings.

One person carries the vision. One person builds the product. One person owns operations or growth. Beyond that, you’re just adding politics.

A Co-Founder Is Not Just an Early Employee

The test is simple: if this person left tomorrow, could you replace them?

If the answer is yes, they are a first hire, not a co-founder. A co-founder is someone whose absence would require you to fundamentally rebuild what you’re doing, not just fill a role but rethink the foundation. Not just someone who was there early, or works hard, or shares your energy. Someone structurally essential.

When Fahim picked me for HackHouse, that was the test he was applying even if neither of us named it. Without a technical, local co-founder, everything he wanted to build was theoretical. I wasn’t nice to have. I was load-bearing.

Get that clear before you hand out the title. Get it clear before you hand out the equity. Because undoing it later is expensive, painful, and sometimes fatal.

Before You Commit, Answer These

Use these five questions as a filter before you make it official.

One: Can you clearly name two or three things this company cannot survive without that you personally cannot provide? If you can’t name the gap, you’re not ready to fill it.

Two: Have you worked through at least one real conflict with this person and come out the other side without permanent damage? Harmony under good conditions means nothing. You need data from a bad day.

Three: Do you both know what the other actually wants from this, financially and personally? Not what you assume. What they’ve told you directly.

Four: If this person disappeared tomorrow, would you have to rebuild the company or just hire a replacement? One of those answers makes them a co-founder. The other makes them employee number one.

Five: Can you sit across from this person for ten hours in a bad week and still respect them? Because that week is coming. Multiple times. Make sure the answer is yes before you’re already inside it.

The first few months at Pathao, nobody had a title. We didn’t have business cards or org charts or reporting lines. We had a shared Google Sheet, three laptops, and an overwhelming conviction that we were going to figure this out. Adnan was writing code. Riaz bhai was on the ground talking to riders. I was doing everything else, which meant I was doing nothing particularly well.

It didn’t take long to realize the problem. Not that we lacked skill, but that we lacked clarity. When you don’t define who owns what, nobody truly owns anything. The product had bugs that nobody was officially responsible for fixing. The driver onboarding was broken because nobody had formally claimed it. We were all vaguely responsible for everything, which is another way of saying we were each fully responsible for nothing. Three people running in slightly different directions is just as bad as running in no direction at all.

So we had the conversation. Not about titles, but about territory.

Titles Are Costumes. Accountability Is the Real Job.

Here’s what nobody tells you about a three-person founding team: the org chart doesn’t matter. Nobody cares that you’re the “Chief Operations Officer” when there are exactly three people in the company and one of them is also handling vendor negotiations. What matters is the answer to a much simpler question: who is going to make sure this specific thing doesn’t fall apart?

That’s it. That’s the whole framework.

At Pathao, we landed on something loose but functional. Adnan owned the product and the technology. What lived inside the app was his domain. Riaz bhai owned the supply side, driver acquisition, and street-level operations. I owned everything facing outward, investors, partnerships, strategy, and the general chaos of keeping the company moving forward.

It wasn’t perfect. It overlapped constantly. But it was clear enough that when something broke, we knew who the first phone call went to. That clarity alone changed how we worked.

Ownership Is a Posture.

Here’s where most founding teams get it wrong. They confuse having a role with having a limit. They assume that because Adnan owned the product, I didn’t need to understand it. Or that because I handled investors, Riaz bhai didn’t need to care about the pitch narrative.

Wrong. Dangerously wrong.

Every founder needs to own everything, even the things that aren’t theirs to execute. This isn’t about micromanaging your co-founders. It’s about the psychological posture you bring to the company every day. When a driver complained that the app crashed during peak hours, that was Adnan’s problem to fix, but it was also my problem, because it affected retention, which affected our investor metrics, which affected our runway. Ownership without boundaries is what separates founders from employees.

This posture gets tested hardest when the company is under physical stress, not just strategic stress.

During one of our early delivery surges, our fleet was completely overwhelmed. Orders had spiked, riders had called in, and there was simply nobody left to run the remaining parcels. I had a meeting in the afternoon and a call with a potential partner in the evening. I cancelled both. I put on a backpack, loaded it with packages, and spent the next several hours delivering around Dhaka on a bicycle.

I want to be precise about how that felt, because the memory is still sharp. It wasn’t heroic. It was hot, sweaty, and the addresses were hard to find. One customer kept calling me to ask where I was, as if my description of “two minutes away” somehow made traffic disappear. Another one wasn’t home after I’d ridden thirty minutes to reach them. I sat outside their building and waited.

Here’s what that afternoon gave me that no strategy meeting ever could: I understood, viscerally, what our riders dealt with every single day. The bad addresses. The customers who didn’t pick up. The weight of the bag on your back after four hours. When I got back and started thinking about onboarding improvements, about fleet incentives, about why our completion rates were lower than they should be, I was thinking from the inside out. Not from a dashboard.

No job is too small. That’s not a motivational poster. It’s a rule of survival. When the company needs something done and nobody else can do it, you do it. The moment a founder treats certain work as beneath them is the moment they start losing touch with the company they’re building.

Builders and Sellers: Know Which One You Are

There are two fundamental jobs in every early-stage company. Building the thing, and convincing the world the thing matters. Everything else is support work for one of those two functions.

Understanding which one you are, naturally, is the first act of accountability. Not because you’ll only ever do one, but because knowing your default wiring helps you recognize when you’re operating outside it, and why that’s sometimes exactly what the company needs.

Here’s a rough test:

You’re probably a Builder if:

• You get more energized shipping a feature than closing a deal

• Your instinct when something breaks is to fix it yourself

• You think in systems, flows, and edge cases

• Ambiguity energizes you when it’s technical, exhausts you when it’s social

• You’d rather show someone a working product than describe it

You’re probably a Seller if:

• You get more energized from a handshake than a deployment

• Your instinct when something breaks is to manage the narrative around it

• You think in relationships, incentives, and timing

• You can walk into a room of strangers and leave with commitments

• You’d rather describe a vision so clearly that someone can feel it

But here’s the problem: in the early stage, you don’t get to just be one. You have to be both, badly and simultaneously, until you can afford to hire the version you’re not.

I wrote code I’m not proud of. Adnan sat in merchant meetings he found excruciating. You do what the company needs, not what your job description says, because in a three-person startup, the job description is just “whatever isn’t getting done.”

The goal isn’t to become equally good at building and selling. The goal is to be self-aware enough to know which hat you’re wearing, and honest enough to know when to hand it off.

When Nobody Owns It, It Dies

About four months into Pathao’s operations, our driver onboarding was a mess. Drop rates were high. Drivers were arriving for registration and leaving confused. The process required paperwork we hadn’t fully thought through, and there were steps in the app flow that made no sense if you’d never owned a smartphone before.

Everyone knew it was broken. Nobody owned fixing it.

The problem lived in the gap between Adnan’s product work and Riaz bhai’s field operations. Each of them assumed the other had it. I assumed they had it. So it just sat there, bleeding quietly, costing us drivers we’d spent real money to acquire.

When we finally forced the question, who owns onboarding completion rates, the whole thing got fixed in two weeks. Not because we suddenly became smarter. Because there was now a person who went to sleep thinking about it and woke up accountable for it.

That’s the lesson. It sounds almost embarrassingly simple. But in the fog of early startup life, accountability gaps are where companies quietly die. Not in the big dramatic decisions, but in the unclaimed territory between roles.

Own your lane. And then own the lane next to it too. And when the fleet is overwhelmed and someone needs to strap on a backpack and go, own that too.

There is a book that every founder in Silicon Valley eventually reads. It’s called Radical Candor, written by Kim Scott, a former Google and Apple executive who spent years thinking about why some managers build extraordinary teams and others build ordinary ones. Her answer is deceptively simple: care personally, challenge directly. Say the hard thing, say it to the person’s face, say it because you actually give a damn about them. The framework is clean. The logic is airtight. In a conference room in San Francisco, surrounded by people who went to the same schools and speak the same professional language and share a cultural assumption that feedback is a gift, it probably works beautifully.

Then you try it in Dhaka.

The Room That Didn’t Want to Hear It

We had a product manager in the early Pathao days who was sharp. Quick-thinking, technically literate, genuinely passionate about what we were building. I liked him. The team liked him. He showed up early. He stayed late. He volunteered for everything.

He was also, quietly, killing the team around him.

Not through malice, but through certainty. He was the kind of person who had an answer for everything before you’d finished the question. When someone proposed a different approach, he’d pause just long enough to make it look like he was considering it, then explain, with complete calm and absolute confidence, why his original idea was still correct. Feedback bounced off him like raindrops off a car roof.

I saw the friction it was creating. I watched two engineers stop raising objections in his sprint reviews. I noticed that one of our best product designers had started going quiet in meetings where she used to be the loudest voice in the room.

For two months, I said nothing directly. I had pulled him aside and hinted. I had framed things as “team dynamics.” I had used every diplomatic trick I knew. Nothing landed. Because nothing I said was clear enough to be heard.

When I finally sat down and told him exactly what the problem was, specifically and without cushioning, he went silent. Then he argued. Then he came back the next day having processed it. And then, slowly, he started to change.

The conversation I should have had in week two happened in month three. We lost sixty days of team performance because I didn’t want to make it uncomfortable.

Why Bangladesh Makes This Harder

In Bangladesh, feedback is personal. That is not a criticism of Bangladeshi culture. It is just the truth of how most of us grew up. In school, a teacher correcting you in front of the class was not considered coaching. It was humiliation. At home, criticism from an elder landed as judgment, not guidance. In the workplace, a manager pointing out a flaw in your work is often interpreted as a verdict on your worth as a person, not an observation about a specific decision.

This is compounded by hierarchy. Bangladesh runs on it. There is a deeply embedded assumption that seniority is tied to correctness. A junior employee challenging their manager’s thinking isn’t seen as intellectual courage. It reads as insubordination, or at the very least, bad manners. And ego, once engaged, becomes load-bearing. People would rather be wrong quietly than right in a way that publicly embarrasses someone above them.

In Silicon Valley, the cultural assumption is that ideas are separate from identity. You can attack the idea without attacking the person. That separation is almost entirely a learned, cultural construct. It didn’t arrive naturally. It was built, deliberately, over decades of tech culture that incentivized shipping over ego, iteration over perfection.

We didn’t have that foundation in Dhaka. We had to build it from scratch, in the middle of trying to build a company.

No Feedback Loop, No Growth

Here is the thing, though. The cultural difficulty of giving and receiving feedback doesn’t make feedback optional. It makes it more expensive when it’s missing.

A startup without honest feedback loops is not a startup. It is a slow-motion disaster that hasn’t announced itself yet. Bad product decisions don’t get caught early. Poor performers don’t get corrected. Good people watch problems fester and start quietly updating their CVs. The team dynamics that should be your competitive advantage become the thing that slows you down.

At Pathao, we learned this the hard way in multiple directions. There were times when we moved forward with a feature that three engineers already knew was wrong but didn’t feel safe saying so directly. There were product bets we doubled down on because no one wanted to be the person who told leadership the data was pointing somewhere else. There were managers who were creating quiet toxicity in their teams, and we didn’t find out until two or three people had already left.

Netflix built one of the most aggressively feedback-dense cultures in tech history. Their foundational belief was simple: if you only hire adults and treat them like adults, they can handle hearing the truth. Every project got a post-mortem. Every manager was rated by their direct reports. Silence was not considered polite. It was considered wasteful. Amazon went even further, structuring meetings so that memos were read in silence first, forcing real engagement with ideas before any social dynamics could take over. The goal in both cases was identical: strip out the noise so that the signal gets through cleanly.

We couldn’t copy those playbooks wholesale. But the underlying logic was undeniable. Fast companies need fast feedback. And the only way feedback is fast is if giving it and receiving it doesn’t require an act of personal courage every single time.

Building the Habit, Not Just the Speech

What actually changed things at Pathao wasn’t a policy. It was a set of repeated behaviors that slowly rewired what people expected from each other.

We started treating public praise as specific rather than general. “Good job” means nothing. “The way you restructured that driver payment flow cut our support tickets by 18% in two weeks” means everything. Specific praise sets a standard. It tells the team what excellence actually looks like, which makes it easier to name what falls short of it.

We started making feedback a two-way street in one-on-ones. Not just me evaluating someone, but explicitly asking: what am I doing that makes your job harder? At first the answers were polite and empty. Then, slowly, they got real. Someone told me I was making too many decisions in sprint reviews without hearing the team first. They were right. I changed it. And because I changed it visibly, other people started believing the feedback was real, that it actually went somewhere.

We also started normalizing mistakes in group settings, not with shame, but with storytelling. When something went wrong on a feature launch or an ops decision, we’d debrief it publicly, not to assign blame, but to extract the lesson before it calcified into unspoken fear. It felt uncomfortable the first few times. Then it started to feel normal. And when it felt normal, people stopped hiding problems.

The Actual Cost of Avoiding the Conversation

There is a version of kindness that is really just cowardice dressed up in good intentions. Avoiding a hard conversation because you don’t want to upset someone is not protecting them. It is protecting yourself from discomfort while letting them continue in the wrong direction. By the time you finally say something, the gap between where they are and where they need to be is so wide that the conversation lands as a verdict rather than a course correction.

Startups grow people or they lose people. There is no third option. The engineers, product managers, operations leads who joined Pathao in its first two years had more professional growth packed into twenty-four months than most people get in a decade of corporate employment. Not because we had the best training programs. Because we had the kind of environment where things moved fast enough that feedback had to be real, had to be quick, and had to be absorbed immediately or it was useless.

That environment didn’t happen by accident. It had to be built. It had to be modeled by leadership first. And it started, every single time, with someone willing to have the conversation that nobody else wanted to have.

The hardest one, the one that feels like it might break something, is usually the one that was already three months overdue.

Here’s the rule. Four years, one-year cliff. Every founder. Every early employee. No exceptions.

It’s not a Silicon Valley quirk or a VC formality. It’s the single most protective structure a startup can have, and it’s been standard for decades because every generation of founders learns the same lesson the hard way. You build with people who believe in the vision. Then time passes. Priorities shift. Someone leaves. And suddenly the question of who owns what becomes the most expensive conversation you’ve ever had.

Why Four Years, Why One Year

The four-year timeline isn’t arbitrary. It roughly mirrors the arc of a startup. Year one, you’re figuring out if the idea works. Year two, you’re building toward product-market fit. Year three, you’re scaling. Year four, you’re either a real business or you’ve admitted it isn’t. By the end of that cycle, each founder’s contribution to the outcome is mostly visible. The equity they hold should reflect the work they actually did, not the bet they made on a Tuesday night when everyone still believed they’d be together forever.

The one-year cliff is the most misunderstood piece. Founders hear it and think it’s punitive. It isn’t. It’s the most humane part of the structure.

If someone joins the founding team and six months in the fit clearly isn’t working, their vision has diverged, their output doesn’t match their stake, you can part ways cleanly. No shares change hands. No lawyers. No buyback negotiation. You separate, and both of you move forward with dignity.

Without the cliff, that same exit becomes a siege. The person has been vesting monthly since day one. They own a real piece of something. Now you have to buy them out, which costs money you probably don’t have, or let them keep it, which means someone who left at month eight retains meaningful standing in a company they barely built. Neither outcome is clean.

The Handshake Problem

Most Bangladeshi startups handle equity with a handshake. Maybe a WhatsApp message. An understanding that exists clearly in everyone’s head but nowhere on paper. And for a while it works, because everyone is too busy to have a dispute.

But startups don’t stay calm. They hit pressure, money, success. Every one of those things changes what people believe they’re owed.

This isn’t a Dhaka problem. It’s a founder problem. When Chesky and his co-founders were renting air mattresses out of a San Francisco apartment, they were running on informal trust too. But they formalized equity early because the company needed a structure that could outlast any individual’s commitment to it. The equity table has to survive the people on it.

Bangladesh adds a cultural layer. Business here runs on relationships. Drawing up documents that plan for someone’s departure feels like an insult to the person sitting across from you. But the discomfort of that conversation is a fraction of the cost of having it later, when the company is under pressure and there is real money at stake.

When a Founder Leaves Early

Fahim was one of Pathao’s co-founders. He was electric, sharp, connected, and he believed in Bangladesh as a market when almost no serious investor outside the country did. He helped us raise our early money. His energy was woven into what we were building.

Two years in, he left. He had an opportunity to start a new ride-sharing business in Nigeria called Gokada, and he took it. That’s a legitimate decision. Ambition is not a character flaw.

But here’s the question it forced: should Fahim receive the same equity as a founder who stayed for four years? Should someone who built the first half of the company walk away with the same stake as the people who survived the near-death moments, the layoffs, the competitive wars with Uber, the months when payroll was a question mark?

The honest answer is no. It’s not fair to the people who stayed. And without a proper vesting structure, “fair” becomes a conversation nobody wants to have, one where the person who left feels entitled and the people who stayed feel robbed, and both sides are technically right because nothing was written down.

Vesting resolves this before it becomes a conflict. Fahim’s equity vests over four years. He leaves at two. He walks with roughly half his grant, and the unvested remainder stays with the company, available to the people who carried the weight he left behind. No argument. The structure already decided.

Equal Is Not Fair

When you’re founding a company with people you respect, splitting everything equally feels right. It says: we’re a team, nobody matters more than anyone else.

But founders are not equal. Not in time invested, not in risk taken, not in what the company actually needs across four years. Equal splits are an emotional decision dressed as a fair one. Emotions change. Numbers don’t.

In Bangladesh, unequal splits feel like a status statement. Giving someone 15% while you hold 40% says, out loud, that you matter more. That conversation is hard in a culture where face-saving shapes every professional relationship. But the resentment of an equal split that feels wrong grows quietly and finds its way out through the company.

Before the Paperwork

Before you open a legal document, answer these questions with your co-founders out loud: Who is actually driving this? What does each of us look like in year four? What happens if someone needs to leave? What happens if someone stops performing?

The conversation is uncomfortable. Do it anyway. The structure only works if everyone understands why it exists, not just what it says.

Vesting protects the company from human nature. The conversation protects the humans.

The paper matters. But what it’s really doing is making a promise: that the people who stay and build will own what they earn, and nothing more.

The first person I hired at Pathao was a seventeen-year-old kid I met at a party.

Not a startup networking event. Not a campus recruitment drive. An actual party (someone’s birthday, I think) where I was nursing a Coke in the corner and this kid walks up to me and says, “You’re that guy building the motorcycle app, right?”

I was that guy. Still am. But back then, being “that guy” meant something different. It meant you were either crazy or broke. Or both.

“Yeah,” I said. “Why?”

“I’ve been following what you’re doing. It’s brilliant. When can I start working with you?”

I almost laughed. Here’s this high school student, probably hasn’t even finished his A Levels, asking me for a job. But something about the way he said it, not desperate, not naive, just direct. Confident. Like he already knew something I didn’t.

“What can you do?” I asked.

“Whatever you need me to do.”

That was it. That was the interview.

Six months later, he was running product. Making decisions I wouldn’t have trusted to people twice his age. He didn’t wait for permission or job descriptions or org charts. He just saw problems and fixed them.

That’s when I learned the first rule of hiring in a place like Bangladesh: experience is overrated. Hunger isn’t.

The Experience Trap

In Silicon Valley, hiring is easy. Well, easier. Reid Hoffman hops from PayPal to LinkedIn. The Collison brothers leave MIT to build Stripe. Engineers move from Google to Facebook to Uber, carrying their playbooks like inherited wealth. The ecosystem feeds itself.

In Dhaka circa 2015? We had none of that.

No one had worked at a ride-sharing company because ride-sharing companies didn’t exist. No one had built real-time dispatch systems because no one needed real-time dispatch systems. No one understood growth metrics or user acquisition or marketplace dynamics because those weren’t problems anyone had tried to solve.

So when we did find people with “relevant” experience, they often worked against us.

The logistics guy from a traditional courier company would walk in with processes designed for a world where customers waited two weeks for delivery. Forms. Approvals. Physical receipts. He thought in terms of days and weeks. We needed people who thought in terms of seconds and minutes.

The software engineer from a bank would design systems assuming users had infinite patience. After all, banking software has never been about speed, it’s about not losing money. Our users would abandon the app if it took more than three taps to book a ride.

Experience, I realized, could be a liability. These professionals knew how to solve yesterday’s problems. We were trying to invent tomorrow’s solutions.

Finding Hunger in Strange Places

Instead of experience, we looked for something harder to quantify: hunger. Raw, unfiltered drive.

The best people we found weren’t posting on job boards. They were high school students who spent their evenings teaching themselves to code. University kids who showed up to startup meetups even though they had no business being there. People who sent cold messages on Facebook saying, “I don’t know what you do, but I want to help.”

Brian Chesky talks about how Airbnb’s early employees had to be “missionaries, not mercenaries.” Same principle, different context. We needed people who believed in what we were building, not people who just wanted a paycheck.

But finding them meant fighting cultural expectations. Parents wanted their kids in “stable” jobs at banks or multinationals. Friends questioned why anyone would join a motorcycle taxi company when they could work at Grameenphone or British American Tobacco.

The concept of equity was foreign. Stock options? Most people had never heard the term. We spent as much time explaining startup compensation as explaining the actual job.

Working Sessions Over Interviews

Traditional interviews were worse than useless, they were actively misleading. Someone could nail every behavioral question but crumble when faced with real ambiguity.

So we stopped interviewing and started working.

Give someone a project. See how they approach it. Do they ask smart questions? Do they come back with ideas you hadn’t considered? Do they take ownership of the outcome, or just complete the task?

Sometimes two hours told you everything. Sometimes it took two weeks. But working together revealed things no interview could: how they handled frustration, how they communicated under pressure, how they learned from mistakes.

Spotify does this with their “hack weeks”—they throw potential hires into actual projects with real teams. We couldn’t afford hack weeks, but we could afford small projects. Build a simple landing page. Analyze our customer support data. Figure out why driver signups are dropping in Uttara.

The seventeen-year-old from the party? His first project was calling fifty drivers to find out why they weren’t coming online during peak hours. He came back three days later with a spreadsheet, a presentation, and three specific recommendations that increased our active driver rate by 20%.

But I didn’t just throw him the task and hope for the best. I spent hours with that kid. Not just explaining what to ask drivers, but how we thought about problems, how we made decisions, how we treated people who depended on us for their livelihood.

It wasn’t charity. It was math. If I spent twelve hours training him properly, and that training improved his performance by even 1%, I’d gain hundreds of hours of better work over the next year. That’s a 20x return on time invested.

More importantly, I was the only one who could do it effectively. You can’t outsource the transmission of company culture and standards to someone who’s never lived them. The person doing the training has to be a credible role model, someone who’s actually solved the problems they’re teaching others to solve.

That’s when I knew he was different.

The Speed vs. Quality Trade-off

As we grew, every decision became a trade-off between doing things right and doing things fast. We almost always chose fast.

Building an automated compliance system for driver verification would take ten engineers three months. Training twenty people to process documents manually took two hours. Our backend wasn’t sophisticated software, it was an Excel sheet holding the fate of thousands of drivers.

We were a tech company solving problems like a call center. And for a while, that was exactly what we needed to be.

The alternative—building perfect systems while competitors ate our lunch—would have killed us. Uber was coming. Local competitors were launching. We had maybe six months to establish market presence before it became impossible.

Travis Kalanick used to say, “Uber is a math problem wrapped in a political problem wrapped in a human problem.” In Bangladesh, it was that times ten. We couldn’t afford to solve for elegance. We had to solve for survival.

The Mass Hiring Factory

By 2017, we weren’t hiring individuals—we were running a hiring factory. Twenty new people every week. I stopped recognizing faces in the elevator.

This is where all startup hiring advice breaks down. You can’t do careful cultural fit assessments when you’re hiring eighty people a month. You can’t have founders personally vet everyone when entire departments are materializing overnight.

Your early hires become your hiring filters. Your company culture becomes your screening mechanism. People who thrive in chaos self-select in. Those who need structure and predictability filter out naturally.

The challenge wasn’t just quantity—it was maintaining quality while scaling like a rocket. Some weeks we’d hire future leaders. Other weeks we’d hire people who lasted two weeks before realizing startup life wasn’t for them.

Manufacturing Teams from Scratch

When we realized we needed a data science team, Dhaka had maybe five people who could credibly call themselves data scientists. None of them were available.

So I went back to my university, found a statistics professor, and asked for help. “Can you build me a team?”

He walked in a week later with his four best students. They weren’t data scientists, but they were smart, curious, and eager to learn. They grew with us, evolving from solving academic problems to solving real-world problems at scale.

This became our template: identify the skills we needed, find the closest academic equivalent, convert students into professionals. It wasn’t efficient, but it worked.

Stripe did something similar in their early days, they hired people who were smart generalists rather than specialists. Patrick and John Collison figured it was easier to teach smart people about payments than to teach payments experts to be smart.

The LinkedIn Strategy

While hunting for employees, we hunted for mentors. Bangladesh might not have had startup veterans, but it had non-resident Bangladeshis who’d seen how things worked in mature markets.

LinkedIn became our secret weapon. I’d spend hours searching for Bangladeshis at tech companies in Silicon Valley, London, Singapore. The pitch was simple: “We’re building something unprecedented in Bangladesh. Can you help?”

But simple doesn’t mean easy. That outreach message took me twelve drafts to get right:

“Hi [Name], I’m [Your Name], founder of Pathao, Bangladesh’s first motorcycle ridesharing platform. I found your profile and saw you work at [Company] in [City]. We’re solving transportation problems for 20 million people in Dhaka, but we’re doing it without a playbook—because no one’s built this kind of business here before. I’d love 15 minutes of your time to get your perspective on [specific challenge related to their expertise]. We can’t pay consulting fees, but we can offer a small equity stake if you’re interested in helping build Bangladesh’s first unicorn. Would you be open to a brief call?”

The response rate was about 30%, way higher than I expected. But the conversations taught me something crucial: people wanted to help, but they wanted to know exactly how their specific expertise could make a difference. Generic asks for “advice” got ignored. Specific asks about marketplace dynamics, or payment processing, or growth metrics got responses.

Building an advisory network of NRBs gave us more than advice—it gave us credibility. When potential hires saw recognizable names associated with the company, it signaled we were serious.

Equity as Education

We were broke most of the time, which meant we couldn’t compete on salary. But we had something potentially more valuable: ownership.

Your equity is worthless until you have an exit, so why be stingy? We gave equity generously - to employees, advisors, even people who helped us for a few weeks during critical moments.

But first, we had to educate people about what equity meant. Most had never heard of stock options. The concept of owning a piece of a company’s future value was foreign.

This wasn’t just compensation; it was alignment. When someone owns a piece of the outcome, they think like owners, not employees. They care about long-term success, not just monthly salaries.

We gave equity to professors who helped us recruit students. To advisors who spent hours helping us avoid expensive mistakes. To the statistics teacher who built our first data science team. If someone could help us avoid a hundred-thousand-taka error, why wouldn’t we give them a piece of something that was currently worth zero but could become worth millions?

The People vs. Technology Paradox

Here’s the dirty secret: we had armies of people manually processing what should have been automated.

Our compliance team was entirely human-powered. Customer service was agents on phones, not chatbots. Our operations team used WhatsApp groups to coordinate with drivers because building a proper fleet management system would take months we didn’t have.

We were supposed to be a tech company, but we solved most problems the way companies solved them in 1980. And for a while, that was exactly the right choice.

Because in a market moving as fast as ours, with competition as fierce as ours, perfect is the enemy of good enough. The companies that survived weren’t the ones with the best technology, they were the ones that moved fastest.

Amazon started as a bookstore run out of Jeff Bezos’s garage. They didn’t build sophisticated algorithms on day one. They built a website that could sell books, then figured out the rest.

Creating an Ecosystem

We weren’t just filling positions, we were creating entire categories of professionals that didn’t exist before. Product managers who understood marketplace dynamics. Engineers who could build for real-time systems. Operations people who could think in terms of network effects.

By the time we reached scale, we’d directly employed around fifteen hundred people. But more importantly, we’d created a template. Other startups could now hire our alumni. We’d proven you could build world-class teams in markets without existing talent pools.

This is the long game of hiring where talent doesn’t exist. You’re not optimizing for immediate productivity, you’re betting on potential and growth trajectory. You’re creating the foundation for an entire ecosystem.

Some of those bets don’t pay off. People leave. People don’t develop as expected. People get overwhelmed by the pace of change. But the ones who stick become your foundation and proof of concept for everyone who comes after.

When the next wave of Bangladeshi startups started hiring, they could point to our alumni as examples of what was possible. Talent begets talent. Success stories inspire more success stories.

The Proof Is in the Results

Building without existing talent is harder short-term but potentially more rewarding long-term. You get to write the playbook. Set the standards. Create the culture that defines how things get done.

And you get to watch people transform. That seventeen-year-old from the party? He’s now running operations for a logistics company that’s three times the size Pathao ever was. The statistics students who became our data science team? Two of them started their own companies. One joined a Silicon Valley startup.

They didn’t just learn skills, they learned how to learn. How to adapt. How to build things that didn’t exist before.

That’s the real value of hiring for hunger over experience. Experience teaches you what’s been done. Hunger teaches you what’s possible.

Because in the end, talent isn’t found. Talent is created. And when you’re building where no one has built before, creation is your only option.

There’s a question every investor asks early in a meeting. You’ve barely sat down, still adjusting the laptop, and they lean back with this look, like they already know what you’re about to say, and they ask it. “Who is your customer?”

And almost every founder, everywhere in the world, but especially in Bangladesh, gives the same answer.

“Everyone.”

I’ve heard it in Dhaka. I’ve heard it in Singapore. In rooms with chandeliers and rooms with bad fluorescent lighting. Founders say it with complete confidence, like it’s a strength. And the investor nods, writes something in their notebook, and the meeting dies right there, even if neither side knows it yet.

I know this because I was one of those founders. And I know what it costs.

A couple of years ago, I was having breakfast with a French billionaire at the St. Regis in Singapore. Holdings in fintech, real estate, energy, biotech. The kind of person who doesn’t measure opportunity in ideas, but in scalable infrastructure. Over eggs and coffee, he asked me how big my market was. I told him about the population, the GDP growth, the mobile internet explosion. All the things you put on slide three. He listened. Then he put down his coffee.

“That’s not a market. That’s a country.”

He wasn’t being dismissive. He was being precise. A country is not a market. A market is a specific group of people with a specific problem who have the money and motivation to pay for a specific solution. Confusing the two is how founders build products for 170 million people that only 50,000 ever use.

What he was really asking me about was TAM, SAM, and SOM. Three layers of market reality that every founder should know cold, not because investors want to hear them, but because getting this wrong is how you build the wrong product for the wrong people and run out of money wondering what happened.

TAM: The Number That Lies to You

TAM, the Total Addressable Market, is the biggest number. The entire universe of demand if you somehow captured every customer who could possibly use your product. If you’re in ride-hailing, TAM is every person alive who sometimes needs to get somewhere. The number is enormous and completely useless for actually running your business.

TAM tells you whether an industry is worth entering at all. But TAM being large doesn’t mean your opportunity is large. It just means the category has room. You still have to fight your way to a piece of it.

WeWork understood this the hard way. Their TAM pitch was seductive: hundreds of millions of office workers globally, a multi-trillion dollar commercial real estate market, the entire future of work. Investors poured in $47 billion at peak valuation. But strip away the narrative and the SAM was far thinner. Flexible co-working appealed to a narrow slice of freelancers, early-stage startups, and remote workers in premium urban markets. That’s not a bad business. It was never a $47 billion business. When valuation met reality during the 2019 IPO attempt, the whole thing collapsed. The French billionaire would have asked one question and seen it coming: “But who is actually paying, and how many of them are there?”

SAM: Where Honesty Starts

SAM, the Serviceable Addressable Market, is where honesty starts. It’s the slice of TAM you could theoretically reach given your geography, product capabilities, price point, and distribution. What language is my product in? What device does it require? What income level does it assume? What cultural behaviors does it depend on? Every filter removes a chunk of your TAM, and founders hate this because watching the number shrink feels like losing.

It isn’t losing. It’s locating yourself.

In Bangladesh, the SAM calculation is particularly brutal if you’re honest. The headline is seductive: 170 million people, one of the youngest populations in Asia, smartphone adoption exploding. You put those numbers in a deck and investors lean forward. But run the filter and the picture changes quickly. How many can afford to pay for a tech service regularly? How many have consistent data access? How many are in urban areas where digital commerce functions? How many have a bank account or mobile wallet?

Run those numbers and 170 million starts looking like 5 million. In Dhaka specifically, a city of 20 million that every founder uses as their core market, the actual addressable segment for a consumer tech product sits somewhere between 150,000 and 200,000 people. Not 20 million. Not even 2 million.

That morning at the St. Regis, the French billionaire already knew this. “Of those 170 million people,” he asked, “how many earn more than $10,000 a year?” I did the math in my head. Maybe 5 million. He nodded. “Those are your customers. The people ordering food delivery three times a week. Using ride-hailing daily. Complaining about traffic from air-conditioned cars. That’s your market.” Then he said something that landed harder: “In India, that same segment is 100 million people. Twenty times larger. So why would I invest in Bangladesh when I can invest there?”

I didn’t have an answer.

SOM: The Number That Actually Runs Your Business

SOM, the Serviceable Obtainable Market, goes further. It’s the share of your SAM you can realistically capture given competition, resources, and time. If your SAM is 5 million and you’re a seed-stage company with 15 people, you’re not getting 5 million customers anytime soon. SOM is the planning number. The one that makes your financial model real instead of theoretical.

The mistake isn’t misunderstanding these terms. Most founders can define them. The mistake is in the calculation. TAM is easy because it’s just math. SAM requires honesty. SOM gets skipped entirely because it’s uncomfortable.

How the Best Companies Started Small on Purpose

Airbnb’s initial target wasn’t everyone who travels. It was attendees of the 2008 Democratic National Convention in Denver, where hotels were sold out and thousands of people needed somewhere to sleep. Their initial SOM was effectively a few hundred people. That absurdly small focus is what let them learn fast enough to eventually reimagine how the world travels.

Uber launched in San Francisco. Not the United States. One city. High-income professionals with notoriously bad taxi access, smartphones, and credit cards. One customer type, one problem. Expansion came after the model was proven.

Amazon started with books. One category, one customer type: people who read enough that local bookstore selection was limiting. Bezos identified a structural advantage online, served that customer with precision, then used the machine he built to expand into everything else. The constraint was the strategy.

The Bangladesh Trap: Why “Going Big” Is the Wrong Instinct

There is a particular kind of ambition in Bangladeshi business culture that I both admire and find dangerous. The ambition of a people who have survived floods, famines, partition, and political chaos and still built one of the fastest-growing economies in Asia. The instinct is to go big. To reach everyone. Saying “my customer is 200,000 people” sounds like small thinking in a country that has fought its entire existence to be taken seriously.

But vision and market are not the same thing. Trying to serve everyone on day one doesn’t make you ambitious. It makes you unfocused. Unfocused companies burn through money without building anything durable.

There’s also a behavioral layer most founders miss. Unlike San Francisco, where early adopters self-select as people who buy new things for the status of being first, Bangladesh has a rigid social structure around consumption. The upper-middle class, maybe 3 to 5 percent of the population, behaves like tech consumers. Below that, behavior changes completely. People rely on cash. They trust relatives more than interfaces. They’re skeptical of anything that requires changing established habits without a referral from someone in their circle.

Your SAM filter in Bangladesh has to include behavior, not just income. You’re not only asking who can afford your product. You’re asking who has the readiness to adopt it. Founders resist this because it forces the kind of honesty that feels like shrinking the dream.

The dream doesn’t shrink. The target does.

The Exercise, and the Question That Ends It

Draw three columns. First: every person who could possibly benefit from your product, no filters. That’s your TAM. Second: apply every realistic constraint, geography, device, income, behavior, language, existing habits. Be ruthless. That’s your SAM. Third: your team size, budget, network, runway. What can you actually capture in the next twelve months? That’s your SOM.

If the SOM terrifies you because it’s small, you’ve done it right. A small, honest SOM means you know exactly who you’re going after. You can build for them, market to them, retain them, and expand outward from that foundation.

The French billionaire wasn’t wrong. He was just doing the math I should have done before sitting down across from him. When he asked how big my market really was, he was asking for the SAM, not the TAM. I gave him the TAM because it sounded better. He saw through it in thirty seconds because he’d heard the same pitch in twenty other countries.

Before your next pitch, before your next feature, write your SOM on a whiteboard. A real number, not a hope. Then defend it to someone who doesn’t believe in your idea. If you can’t hold that number under pressure, you don’t have a market yet.

You have a demographic.

Demographics don’t buy things. People do.

Not with everyone. With someone.

There is a silence after you launch that nobody warns you about.

You’ve built the thing. Tested it. Argued about the logo and the landing page and the pricing. Told everyone it’s ready. And then, on day one, nothing happens.

No signups. No calls. No replies to your emails.

You refresh the analytics dashboard every twenty minutes. You wonder if something is broken. You wonder if you are. Most founders do something catastrophic at this point: they try to fix their marketing instead of fixing their thinking. They run ads, post on social media, write blog posts. They do anything except the one thing that actually works, which is go find a specific human being who has the specific problem you solve, and get in front of them.

Finding your first customers isn’t a marketing problem. It’s a hunting problem.

Two Games. Completely Different Rules.

Before anything else, you need to understand that there are two fundamentally different ways companies acquire customers, and they require opposite instincts, opposite timelines, and opposite definitions of success.

Consumer is volume and speed. Enterprise is relationships and patience.

In consumer, you’re trying to get strangers to try something without ever talking to you. You’re competing for attention in a market where people decide in seconds. Your product has to be immediately legible, immediately useful, immediately worth telling a friend about.

In enterprise, you’re trying to convince a room of cautious professionals to bet their career on your unproven product. There are multiple decision makers. Procurement cycles. Security reviews. Legal. The deal that takes one afternoon to close in consumer takes six months in enterprise, and it falls apart for reasons that have nothing to do with your product.

Getting these confused is one of the most common early-stage mistakes. A founder with a B2B product runs Facebook ads because that’s what Coca-Cola does. A founder with a consumer app tries to land a corporate contract because it sounds like “real revenue.” Both burn time chasing a customer type their product was never designed for.

Pick your game before you pick your tactics.

How Uber Got Its First Drivers

In 2010, Travis Kalanick and Garrett Camp didn’t launch Uber to the public and wait for downloads. They launched to roughly 400 people, friends and tech insiders in San Francisco, by invite only.

That wasn’t humility. That was strategy.

The supply side of a marketplace is always harder to acquire than the demand side, because supply has to exist before demand can. Nobody gets in a car that isn’t there. So before Uber could find riders, they needed drivers, and before they could find drivers, they needed to make the economics compelling enough to convert professional black car operators away from their existing dispatch systems.

They did this by showing up in person. Phone calls. Direct conversations. Handling objections. Signing drivers one at a time. It was tedious and slow and completely unscalable, and it was exactly right for that stage.

Then SXSW 2011. Free rides in Austin, pointed directly at tech-savvy early adopters who would go home and tell ten other people. Not a mass campaign. A targeted seeding operation.

The consumer playbook, when it works, looks like this: find the smallest, most concentrated group of people who acutely feel your problem. Get the product in their hands through any means necessary. Make the experience undeniable. Let them spread it.

Word of mouth isn’t a marketing channel. It’s the product working.

How Salesforce Crossed Over

Salesforce launched in 1999 into a market dominated by Oracle and Siebel, giants with armies of salespeople and decades of relationships. On paper, no chance.

So Marc Benioff didn’t fight on their terms. Instead of chasing the CIO, the one signing big enterprise software checks, he went after the salespeople who actually had to use the software every day and quietly hated it. Mid-market companies. A single geography. One function only: pipeline management.

The product spread team to team without needing a top-down IT mandate. One department would sign up, it would work, they’d tell the team next door. Salesforce didn’t expand through sales cycles. It expanded through reputation. The enterprise version of word of mouth isn’t someone tweeting about your app. It’s an internal champion vouching for you in a budget meeting.

But here’s the sharper lesson: Slack moved the other direction entirely. Stewart Butterfield built Slack initially as an internal tool for his own gaming company, a consumer-like product used by a small, specific team. It spread inside that company first. Then it spread to other tech companies, initially through the same informal, viral mechanism: individuals adopting it, adding colleagues, the product growing through genuine daily use rather than sales. Only after it had proved itself in this way did Slack build an enterprise motion, with security features, admin controls, and corporate contracts.

It started consumer-adjacent and migrated to enterprise. Salesforce started enterprise and eventually built a platform that developers adopted almost like consumers. The direction is less important than the sequence: prove it small, then expand the surface area.

One game transitions into the other. But you can’t skip straight to the second game.

What Rejection Actually Feels Like

Nobody talks about this honestly. So let me.

You will send 200 cold emails and get 4 replies. Two of those will be “not interested.” One will be a referral to a junior person who also won’t respond. The fourth will take three follow-ups to convert into a 20-minute call, on which the person will seem interested, then go silent for six weeks.

And you’ll sit there wondering: is the idea wrong? Is my pitch wrong? Am I wrong?

The silence is not a signal about your idea. It’s a signal about your targeting. Most early outreach fails because founders are emailing the market instead of a person. The message is generic because the audience is generic, and generic messages produce generic results, which is nothing.

The fix is to get uncomfortably specific. Not “companies that could use our software” but “Series A fintech startups in Singapore that hired a head of compliance in the last 90 days.” Not “people interested in fitness” but “women over 35 who signed up for a half marathon for the first time.” The more specific your target, the more specific your message can be. The more specific your message, the higher your reply rate. Every time.

The rejection isn’t feedback about your product. Most of the time, it’s feedback about your aim.

Your First Ten Customers: What To Actually Do

Stop planning. Do these in order.

1. Write down twenty names. People you know, or can reach through one introduction, who have the exact problem you solve. Not “could potentially benefit.” Actively suffering from.

2. Call them, don’t email. A phone call gets ten times the signal of an email. Speak to ten people before you send a single cold outreach.

3. Ask one question only. “What’s the hardest part of solving this problem right now?” Listen. Don’t pitch. The way they answer will rewrite your positioning.

4. Offer to solve it for free. Your first customers aren’t a revenue event. They’re a research event. Trade the solution for their time and honesty.

5. Earn the referral before you earn the check. After each conversation, ask: “Is there anyone else you know dealing with this?” Your second customer is hiding inside your first customer’s network.

6. Watch what they do, not just what they say. Do they actually use the product? Do they log in once and disappear? Do they bring a colleague in without being asked? Behavior tells the truth that courtesy hides.

The goal of your first ten customers is not revenue. It’s conviction. You’re trying to find out whether your problem is real, whether your solution fits, and whether there’s a repeatable path to finding more people like them.

That path is your business. Everything else is decoration.

At some point around 2018, someone opened a board meeting with a slide that said: “Pathao Super App.”

Everyone nodded. Of course. Gojek was doing it. Grab was doing it. WeChat had essentially become a country unto itself in China. The logic felt airtight: we already had the users, we already had the riders, we already had the infrastructure. All we had to do was keep adding services on top. Rides. Food. Courier. Cars. Keep stacking. Keep expanding. Build the everything app.

So we did.

At our peak, Pathao had ten verticals. Ten. Rides. Courier. Cars. Food delivery. Health. Pathao TV, where users could watch content inside a ride-hailing app, as if the thing people desperately needed between booking a bike and reaching their destination was a streaming service. Ten different product lines, each with their own teams, their own roadmaps, their own monthly reviews where someone presented numbers and someone else asked what we were doing to grow them.

The answer was always the same: build more.

How It Starts

It never feels like sprawl from the inside. Every vertical feels like a logical extension of the last.

Courier was first. That was the original Pathao, three riders, a spreadsheet, a woman named Poter Bibi negotiating us down to 60 taka per delivery. Scrappy and real and actually working.

Then we launched rides because our riders had dead hours between deliveries and we needed to sweat those idle assets. That worked too. Students outside BRAC University, five minutes instead of thirty, word spreading through campus Facebook groups without us spending a taka on ads.

Then Uber entered Bangladesh with cars and we needed to fight back, so we launched Pathao Cars. Then the margins on bikes were thin and food delivery had better economics, and HungryNaki and Foodpanda weren’t serving the Star Kabab crowd, so we launched Pathao Food. Then we were in Dhaka and quietly expanded to Chittagong before Uber could get there.

Every decision had a rational explanation. Every expansion was defensible in a board deck.

Then came health. The pitch was that we had a captive audience, people sitting on the back of a bike with nothing to do but stare at their phones, so why not connect them to doctors, clinics, health services? It sounded compelling in the meeting room. In practice, we were a ride-sharing company trying to build healthcare company without clinical expertise, regulatory relationships, medical trust, or the focused attention the category actually demanded. We launched anyway.

Then came Pathao TV. We had an app with a large daily active user base, so why not keep people inside it longer? The problem was we had neither the capacity nor the budget to produce anything worth watching. So we brought in content partners. Those partners had their own incentives and their own definition of quality. Users opened the Pathao app to book a ride. Nobody opened it to watch a show. The data was unambiguous. The usage numbers on Pathao TV were embarrassing. But we kept it alive for months because killing it meant admitting we had wasted the engineering time that built it, time that could have gone toward making our core rides product better, faster, more reliable, more loved.

That is the hidden cost nobody talks about. Not just the cost of building the wrong thing. The cost of not building the right thing with those same hours.

Why Bangladesh Doesn’t Work Like China

Here is what nobody says loudly enough in founder circles: the super app is a Chinese invention, born in a very specific context that almost nowhere else in the world actually replicates.

WeChat works because China had no dominant incumbent apps when smartphones arrived. The infrastructure gap meant WeChat could colonize everything, messaging, payments, commerce, services, all at once, into a single daily habit. People did not already have a banking app, a food app, a messaging app. WeChat became all of those things by default, before alternatives existed, and with the added tailwind of a government-controlled internet that kept Western competitors out entirely.

Gojek succeeded in Indonesia because the ojek, the motorcycle taxi, was already culturally embedded before the app arrived. The trust existed in the physical world before the product did. Gojek layered services onto a relationship that was already habitual. Grab did something similar across Southeast Asia, starting from a core behavior people were doing every single day long before any app told them to — but here’s the secret — 95% of the volume comes from top 3 services.

Even then, truly successful super apps you can count on one hand globally. The companies that tried to copy the model without those underlying conditions mostly burned through capital trying to manufacture the kind of habitual daily use that only comes organically, from solving one problem so completely that people cannot imagine their day without you.

We were trying to do it in Bangladesh, a market of 170 million people that sounds enormous until a French investor over breakfast in Singapore asks you how many of them can actually spend money on your service regularly. The honest answer is around five million. That is your real addressable market, the people who use ride-hailing daily, order food delivery multiple times a week, and have the disposable income to make those habits stick. Five million people. Smaller than the city of Singapore.

And here is the painful irony. Once you understand the addressable market is that narrow, you also understand why you were tempted to build ten verticals in the first place. Because no single product, in a market that size, will ever build a company worth building on its own. You need to stack revenue streams just to survive.

The Conglomerate Lesson Bangladesh Already Knew

The old money in Bangladesh figured this out decades before startups arrived, and they did it without ever calling it a super app.

Look at Bashundhara Group. Cement. Paper. LPG. Real estate. Media. Food. Steel. They are in everything, and that is not an accident or a lack of strategic focus. It is the only rational response to building in a small market. You diversify because any single industry has a ceiling that arrives too quickly. You spread across sectors not to become everything to everyone, but to pool operational resources, shared distribution, shared brand equity, shared management bandwidth, so that each business line does not have to carry the full overhead of running entirely alone.

Pran-RFL is the same story. They manufacture food products, plastic goods, beverages, garments, agro-processing. They export to over 140 countries. The multi-vertical structure was never their weakness. It was always the point. The shared physical infrastructure, the factories, the supply chains, the distribution networks, that is what makes each individual business viable at a scale a standalone company in Bangladesh could never reach on its own.

A conglomerate is, in its own way, a super app. Shared resources. Multiple revenue lines. One holding structure. But here is the crucial distinction that Pathao missed and Bashundhara never confused: conglomerates never told their customers that the reason to engage with them was because they did everything. Bashundhara did not build a single consumer-facing app and ask you to buy cement, order food, read their newspaper, and pay your electricity bill all in one place. They built businesses that happened to share infrastructure on the back end, while presenting each product to the market on its own terms.

We were trying to do the opposite. We were trying to tell users: come to the Pathao app because we do everything. And what users heard was: we’re not sure what we are anymore.

The super app strategy only works when your users have a reason to live inside your product all day. WeChat earned that because it owned social communication first. In Bangladesh, Pathao earned something real too: the daily commute. That is a meaningful relationship. It is not an all-day relationship. And trying to turn a rides relationship into an everything relationship, through health features and streaming content and services people could get elsewhere with two taps, is not strategy. It is wishful thinking dressed up in a board deck.

The Push and Pull No One Prepares You For

So here is the genuine tension that every founder building in a small market eventually hits.

On one side: the market is too small for a single-product company to justify the capital required to build it properly. You need multiple products to hit numbers that satisfy investors, attract senior talent, and give you enough runway to survive the inevitable hard stretches.

On the other side: if you spread across too many verticals without real traction in each one, you end up with a bloated product that confuses users, drains engineering focus, and creates the illusion of a company while the core slowly weakens underneath all the noise.

Both things are true at the same time. And there is no clean resolution. What you are really doing is managing the tension.

The mistake we made was chasing the super app narrative rather than the super app economics. We added verticals because investors liked the story, because Gojek’s valuation kept climbing, because building felt like momentum. We did not ask, for each new vertical, whether we had a genuine right to win in that space. Health required clinical depth we did not have. TV required content production we could not afford to do properly. We launched them anyway, because momentum is seductive, and standing still feels like dying.

What Actually Works: Build Sharp Here, Expand Everywhere

Spotify was founded in Sweden. Population: ten million people. A market so small that by the logic of the super app believers, it should never have produced a global company worth hundreds of billions of dollars.

What Spotify did instead of expanding verticals was obsessively deepen the one thing it had. They did not launch Spotify Health or Spotify TV or Spotify Rides. For years they built a single music streaming experience and they built it until it was so frictionless, so genuinely better than every alternative, that the product itself became the argument for using it. When they finally moved into podcasts, the criticism was loud and immediate. Investors questioned the logic. Journalists wrote about scope creep. Even internally, the move was contested.

But the expansion held because Spotify applied a test that most founders skip: is the user habit the same? Music and podcasts are both about listening. You open the app, put something in your ears, and move through your day. The behavior did not change. The content did. That is an adjacent expansion that reinforces the core relationship rather than replacing it with a new one.

When Spotify eventually moved into audiobooks, same test, same answer. Listening is listening. The habit transferred.

Compare that to Pathao TV. Watching video and booking a motorcycle ride are not the same habit. The user who opens the Pathao app has a specific, time-pressured intention: I need to get somewhere. Putting a content tab in that context is not expanding the relationship. It is confusing it. The user who wants to watch TV is not the same user who needs a bike in eight minutes, and designing a product that tries to serve both in the same moment, in the same app session, is designing a product that does neither well.

The question before any expansion is not “do we have the technical capability to build this?” The question is: “is the underlying user habit identical, or are we asking users to change how they relate to us entirely?” If the answer is the latter, you are not deepening your product. You are starting a new one inside the same app and calling it a strategy.

The real model for small-market founders is more specific still. Use a small market as a laboratory. Build something that the market’s low competition and low operating costs allow you to test and refine without burning through capital fighting well-funded incumbents. Get genuinely good at it. Make it sharp. Make it operationally excellent. And then expand geographically, not into more features, but into more markets where the same product meets the same core habit.

Nepal was our proof of this. We did not take Pathao TV to Kathmandu. We did not take Pathao Health. We took the one thing we actually knew how to operate, bike-based ride-hailing, and we dropped it into a city that looked like Dhaka three years earlier. Narrow roads, gridlocked traffic, a bike culture already embedded in daily life, and no well-funded competitor in sight.

The first few weeks were electric. We had made history, the first Bangladeshi startup to cross a border and set up internationally. Then the taxi cartel found out about us. Thirty drivers surrounded our office while our employees were still inside. They threw rocks through the windows. They followed our riders and rammed their bikes with taxis to make them feel unsafe. Some riders quit out of fear. Some of our own staff handed in their notice.

We could not fight this battle the way we had fought regulatory battles in Dhaka. There, we knew the players, the ministries, the press contacts, the political angles. In Kathmandu we were outsiders. Bangladeshis in a Nepali market, seen as interlopers. If we tried to handle it ourselves we would get crushed.

So we stepped back and let the Nepali team lead. They went to the media, to influencers, to politicians. They framed Pathao as a solution to Kathmandu’s transportation chaos, as employment for young people, as a public good the city needed. The cartel’s intimidation became a story about powerful incumbents trying to strangle competition. The narrative flipped. More riders signed up. More users ditched taxis for bikes. We won.

And what made that possible was precisely the fact that we had not tried to bring ten products into a foreign market. We had brought one product that we understood completely, that we had stress-tested in Dhaka for years, that we knew how to operate under pressure and in chaos and against hostile incumbents. The focus that the multi-vertical strategy had robbed us of in Bangladesh was exactly what saved us in Nepal.

What the Sprawl Actually Cost

I want to be honest about this because founders tend to nod along to the lesson about focus without really feeling the weight of what unfocused expansion costs. The money spent building things nobody uses is visible. The cost underneath it is harder to see.

When Pathao was running ten verticals, our best engineers were not working on making rides better. They were distributed across health features nobody opened, content infrastructure that served no one, and partner integrations that should never have been priorities in the first place. Every sprint, product managers who each believed their vertical was the most urgent thing in the company competed for the same engineering hours. They were not wrong individually. They were collectively producing a company that excelled at nothing.

The core rides product, the thing that had gotten us to a hundred thousand rides a day, the thing users actually loved and depended on and talked about, stopped getting the attention it deserved. Bugs that should have been fixed in a week sat for a month. Features that would have meaningfully improved the booking experience got pushed down the backlog in favor of Pathao TV integrations that served a user base of nearly zero.

And while we were distracted, the competition was not standing still. Every sprint we spent building something nobody asked for was a sprint we did not spend making sure a user who had a bad experience would give us another chance.

That is what unfocused expansion actually costs. Not just the money spent building the wrong thing. The compounded cost of the right thing not getting better, month after month, while the roadmap kept growing and the focus kept fracturing and the users who mattered most quietly started looking at what else was available.

The road has to lead somewhere. Not ten places. Somewhere. Build until your answer to “what are you?” is so clear and undeniable that no investor, no competitor, no internal meeting can muddy it. Then ask what comes next.

There’s a particular kind of meeting that happens in every startup in Bangladesh. Someone in the room, usually the most eager person there, makes the suggestion: “Let’s just go cheaper than everyone else. That’s how we win.”

It sounds logical. You are a new company. Nobody knows you. The fastest path to getting people through the door is to make the price impossible to refuse.

And that logic will quietly kill your company.

Cheap is not a strategy. It is a panic response dressed up as a plan. The moment you compete on price, you have accepted a race you cannot win, because there will always be someone willing to lose more money than you to beat you. Price is not just a number. It is a statement about what category you belong to, what customer you are designed for, and how seriously you take your own product. Get it wrong, and no amount of great execution will save you.

The Signal You Don’t Know You’re Sending

When Netflix launched in 1997, it did not try to out-cheap Blockbuster. Blockbuster had the stores, the catalog, the brand. If Netflix had just been a cheaper version, it would have died in the comparison.

Instead, Netflix charged $15.95 a month for unlimited rentals with no late fees. Not cheap. But precise. It removed a specific pain customers already knew they had. The price was not the reason to sign up. The price was the proof that Netflix understood the problem.

Your pricing communicates what problem you solve. Not in a brochure. Right there, in that number. If you charge too little, you signal that the problem is not that serious, or that you do not fully believe in your solution. In Dhaka especially, founders assume that low prices equals accessibility equals growth. But accessibility without perceived value creates a customer who tries your product once, does not take it seriously, and never comes back.

The Bangladesh Bazaar Problem

There is a cultural dimension to this that nobody talks about honestly. In Bangladesh, negotiation is woven into everything. At Karwan Bazar, nobody pays the first price for hilsa. At New Market, the tailor’s opening quote is a starting position, not a final offer. Bargaining is not rudeness here. It is how trust is established, how both sides signal they are serious. When your grandfather haggled over a piece of land in Narayanganj, the back and forth was the contract.

This creates a brutal problem for any startup that prices a digital product.

Your early customers will apply bazaar logic to your SaaS subscription or your service fee. They will ask for a discount. If you give it, they will ask for another. Somewhere in that spiral, you stop being a product they respect and become a vendor they manage. The price floor collapses, and suddenly you are running a charity with a login page.

I watched this happen to founders I knew personally. A logistics SaaS company, genuinely good product, spent eight months negotiating custom deals with every enterprise client. Every client got a different price. When the founder tried to standardize a year in, customers who had been paying discounted rates felt cheated. Three of them left. Revenue went backwards before it went forward.

The way through this is not to refuse the conversation. It is to redirect it. When someone asks for a discount, the question you answer is not “how low can I go?” It is “what value are you not seeing yet?” A tailor on Mirpur Road who knows his stitching is different does not drop his price when someone haggles. He picks up the hem, shows the finish on the inside seam, and explains why it lasts three years instead of one. The price holds because the value is visible. Your job is to make the value that visible, every time.

The Founder Who Held the Line

I know a founder in Dhaka who built a B2B HR software company. When he launched, the market expectation was for something cheap, practically free, because that was what the two competitors before him had trained the market to expect. Both of those companies had burned out quietly, unable to sustain the economics.

He priced at three times what the market expected. Prospects pushed back immediately. Some walked out of meetings. His own team internally questioned whether the number was too aggressive. For the first four months, he closed almost nothing.

He did not move the price. Instead, he sharpened his pitch. He started calculating and presenting the actual cost of manual HR processes for each prospect, in taka, per month. Time spent, errors made, compliance risks carried. Then he put his software’s monthly fee next to that number.

By month six, he was closing consistently. His customers were not price-sensitive buyers who had reluctantly settled. They were convinced buyers who understood what they were paying for and why. Two years later, his retention rate was above 90 percent. His competitors had come and gone. He was still standing, with margins that let him actually build the product.

His customers chose him because he was expensive. The price told them he was serious.

What Happens When You Have to Raise Your Prices

bKash launched in 2011 trying to solve a genuinely hard problem: moving money in a country where most people had no bank accounts. To get people onto the platform, they made the pricing nearly invisible. Cash-in was free. Sending money was negligible. The friction of cost was removed so that the harder friction of behavior change, getting people to trust a phone with their money, could be the only battle being fought.

It worked. bKash became infrastructure. Then, slowly, the fees came. Cash-out charges increased. Transaction fees appeared across services. The backlash was real and loud. But almost nobody left.

Because by the time bKash raised prices, the switching cost was enormous. Your family sends money on bKash. Your landlord expects bKash. The pharmacy takes bKash. The network was so embedded that the price correction was painful but not fatal.

This is the only responsible version of the “start cheap, raise later” strategy: you must build something so essential that when the price correction comes, the cost of leaving is higher than the cost of staying. The mistake founders make is assuming they can replicate this with a product that lacks bKash’s lock-in. If your product is not becoming infrastructure for your customer, you will never raise the price without losing them. You will have trained them to expect cheap forever.

Free is Not the Same as Cheap

When OpenAI launched ChatGPT for free in late 2022, it hit a hundred million users in two months. That free tier was not charity. It was a distribution strategy. The product itself was the marketing.

But free is not a business. So ChatGPT Plus launched at $20 a month, a deliberate premium signal. Millions subscribed because the free tier had already proven the value. Now OpenAI is exploring advertising for its free users, the same architecture YouTube built: free with ads for the mass market, paid without ads for people who value their time, and creator monetization as a third layer entirely.

This is what sophisticated pricing actually looks like. Not one number chosen in a meeting. A system designed around different levels of willingness to pay. Most founders price their product once at launch and never revisit it. That is not pricing. That is guessing.

Set the price. Defend the price. And treat it like a product, something that ships, gets feedback, and improves.

When Foodpanda and HungryNaki dominated food delivery in Dhaka, we looked at what they were doing and made a decision that felt almost too simple. We were not going to fight them on their turf.

Their turf was Gulshan restaurants, 500-taka salads, Thai fusion, imported sushi. Glossy apps. Formal restaurant partnerships. Weeks of negotiation, contract signing, POS installations, API integrations. They were built for customers who didn’t flinch at delivery fees that cost more than the meal itself.

Our customer wanted Star Kabab. Biryani in a paper box. Tehari ladled out of an aluminum drum the size of a small child. Food that was fast, familiar, and under 200 taka total. No one was serving them. Not seriously. Not with any respect.

So we didn’t build a better Foodpanda. We went the other direction entirely. Our riders walked into restaurants, took photos of laminated menus and handwritten chalkboards, and our interns manually typed them into the system. No contracts. No negotiations. No onboarding. If a restaurant served food, it was on Pathao. When an order came in, the rider walked up like any regular customer, paid, waited, picked it up, and delivered it. The restaurants had no say. They couldn’t stop us. Within weeks, while competitors were still onboarding their fifth premium restaurant, we had thousands of eateries. We were the first delivery app to offer Star Kabab. How do you operate in Dhaka and not offer Star Kabab? That single decision, to focus on a completely different customer rather than compete for the same one, did more for Pathao’s food business than any feature we built or campaign we ran.

But I want to be honest about something, because I’ve lived both sides of this. The instinct when you face competition is not to choose. It’s to do everything. Better app, lower prices, more restaurants, faster delivery, bigger call center. You don’t make that decision because you’re stupid. You make it because the pressure to show growth across every metric is relentless. Investors ask why your average order value is lower than the competitor. Board meetings turn into comparison exercises. And in Bangladesh specifically, there’s a cultural layer on top of all of it: you never want to appear weak. Saying “we’re not going after that segment” sounds like an excuse. Like you can’t. Like you’re afraid. So instead of a deliberate strategic choice, you frame it as “we do everything, we’re just better at this part.” That framing is the trap. It’s not a strategy. It’s insecurity dressed up as ambition.

The result is that you spread so thin you become mediocre everywhere. You’re not winning on price because your costs are too high from doing too much. You’re not winning on quality because you’re moving too fast to fix anything. You’re not winning on focus because you can’t say no. There are five real ways to beat a competitor. Price. Quality. Service. Functionality. Focus. That’s the whole map. Every company that ever won a crowded market picked one or two and went so deep that competitors couldn’t follow without abandoning everything they’d already built. Southwest chose price and redesigned their entire operating model around it. No assigned seats. No frills. Just cheap and on time. When bigger airlines tried to copy them with budget sub-brands, they found out that cheap isn’t a pricing decision, it’s a cultural commitment you either make all the way or not at all. Netflix didn’t beat Blockbuster on everything. They beat them on one thing: no late fees, to your door. That was the wedge. Spotify didn’t out-quality iTunes. They changed the axis entirely, from ownership to access. One idea, total conviction.

We did the same thing with bikes before we ever figured it out with food. When we launched motorcycle rides in Dhaka, we weren’t offering a premium experience. The bikes were basic, the app was rough, the experience was chaotic. What we offered was something nobody else was offering at all: a way to beat Dhaka traffic on two wheels, at a price that actually made sense for the people stuck in it. We weren’t better than Uber. We were pointed at a completely different problem. Pick your axis. Then build everything around that one promise. Your pricing, your hiring, your product decisions, your pitch to restaurant owners with chalkboard menus.

The companies that try to win on all five end up losing on all five. You don’t need to be the best overall. You need to be undeniable in one place.

The first order we ever got at Pathao felt like a miracle. I paced the room, called the rider three times, refreshed my phone until the delivery was confirmed. I told myself: we’ve cracked it.

Then nothing happened for a week.

Orders trickled in ones and twos after that. My riders would message me carefully, almost apologetically: “Bhai, kichu delivery ase?” The spreadsheet barely moved. We had three riders, one in Uttara, one in Dhanmondi, one in Mirpur, and most days they had nothing to do. No dashboards. No algorithms. Just me, a hundred browser tabs, and a prepaid SIM I kept topping up. Pathao, as we had imagined it, was dying quietly.

Here’s what I didn’t understand then: a first order isn’t traction. It’s proof of life. It confirms the mechanics work. It tells you nothing about whether a market actually wants what you’ve built.

The Moment That Rewired Everything

The call that changed everything came from a woman named Poter Bibi. She ran a small clothing store on Facebook and wanted 20 deliveries. Our biggest order yet. She negotiated our 120-taka price down to 60, and I folded immediately because I needed the business more than she needed us.

Then she said something I wasn’t expecting: “I don’t need instant delivery. Just make sure it gets there within two days.”

We had been building a same-day, on-demand service because speed felt like the obvious value. But Poter Bibi didn’t want speed. She wanted reliability. She was one of thousands of Facebook merchants in Bangladesh operating without a functional logistics layer under them. Traditional couriers were slow, unreliable, and built for the wrong era. She just needed someone who would actually show up.

That single sentence exposed a market we’d been standing next to without seeing. Facebook commerce was exploding across Dhaka. Small sellers moving clothes, cosmetics, phone cases, and half of them had no dependable way to get their products to customers. Once we understood that, we stopped pitching on-demand delivery and started calling every online shop we could find. Orders went from one a month to 50 a day.

Airbnb’s founders had a version of this same moment. They thought they were building a cheap place to sleep during conferences. What they discovered, only after doing it manually in New York and sitting with actual hosts and guests, was that they were really selling the feeling of belonging somewhere rather than just staying somewhere. The product they launched wasn’t the product that scaled. The insight that scaled came from staying close enough to customers to hear what they actually valued.

Validation Is Behavioral, Not Verbal

The BRAC University experiment was even more instructive about what validation really is.

We had a small fleet of bikes sitting idle in the afternoons, costing us money between the 11am-4pm delivery window. So we created a Facebook group. No app, no backend, no investor pitch. Just a group, everyone in my contacts, a simple post: “Need a ride? We have bikes. Cheaper than a CNG, faster too.”

Standing outside BRAC in the heat with flyers, pitching students. Most ignored us. One girl asked if the driver would kidnap her. We told her a CNG was more dangerous because you couldn’t jump out. She wasn’t convinced.

But a few tried. When the ride finished in five minutes instead of thirty, something shifted on their faces. They came back the next day. They brought friends. A post went up in a student Facebook group. My phone started ringing from numbers I didn’t recognize: “Are you the guy with the bikes?”

That’s what validation looks like. Not someone saying “great idea.” Not a survey. Not friends being polite. Strangers, with no reason to be generous, choosing your product and then coming back and telling other strangers about it. That’s the signal. Everything before it is just noise.

Uber didn’t validate ride-sharing through market research either. Travis Kalanick and a few friends quietly tested black car rides in San Francisco at New Year’s Eve 2009. No app, no branding. Just a text to a phone number. When those early riders kept asking “when can we use this again?”, that was the proof. The company they then built was answering a question the market had already answered for them.

What You’re Actually Looking For

There is one signal that cuts through everything else: does demand exist without you forcing it?

When riders started showing up at our training center without being recruited, asking how to join, that was the signal. When Gojek noticed us without a formal pitch, that was the signal. When the courier business revived itself during COVID, not through a campaign but because merchants desperately needed it, that was the signal.

The hard truth is that most founders mistake their own excitement for market demand. They launch, get some polite interest from their network, call it validation, and start scaling. Then they discover the demand was borrowed, not built.

Real validation is unglamorous. It happens in a Facebook group. On a campus sidewalk in 35-degree heat. On a phone call with a woman who negotiates your price in half before she’ll even try you. It happens in the gap between what you thought you were building and what the customer tells you they actually need.

The market always tells you the truth. You just have to be humble enough to hear it.

When the investors walked into our Chairmanbari office in early 2016, I watched their faces. I was looking for the moment. The moment where they’d realize we were barely a company. That our “office” was a construction site with wi-fi. That the logo on the wall was held up with double-sided tape.

But the moment never came.

They sat in our bargained-for chairs, nodded at the plywood desks, and dove straight into questions about our model. About our delivery times. About our unit economics. They didn’t care that we looked broke. They cared that we had riders out there, right now, delivering packages across Dhaka.

We weren’t polished. We were operational.

That’s the thing about MVPs nobody tells you. The point isn’t to build something beautiful. It’s to build something that proves you can solve a problem, right now, with whatever you have. And sometimes, learning that lesson costs you five years and a regulatory shutdown.

The $200K That Saved Us

Pathao wasn’t sophisticated when we started taking courier orders. We had no dispatch system. No routing algorithm. No predictive anything. What we had was phones, bikes, Google Sheets, and riders willing to navigate Dhaka’s chaos for 60 taka per run.

Every delivery was manually coordinated. A rider would call saying he couldn’t find the address. We’d call the customer. The customer would describe landmarks. “You know the paan shop with the red awning? Turn left there.” We’d relay it back. The rider would find it. Or he wouldn’t, and we’d start over.

It was absurd. It was inefficient. And it worked.

Because the customers didn’t need perfection. They needed their packages delivered. The merchants didn’t need dashboards. They needed someone reliable to pick up their stock and get it to buyers. We gave them that. Just that. Nothing more.

When those first investors showed up, they didn’t invest in our tech stack or our UI. They invested in proof. We had forty deliveries a day happening in real time. We had merchants paying us. We had riders showing up every morning. That was worth $200,000. Not the potential version of Pathao, the one running on algorithms and venture scale. The actual version. The one held together with phone calls and prayer.

This is what founders miss about fundraising. Investors don’t fund perfect products. They fund evidence. Evidence that you understand a real problem. Evidence that people will pay for a solution. Evidence that you can execute under constraint.

Our plywood desks weren’t a bug. They were a feature. They proved we weren’t burning cash on appearances. Every taka went into riders, operations, and keeping the wheels turning. That told investors something no pitch deck could: we knew how to survive.

When Good Enough Kills You

Here’s where it gets tricky. Shipping fast doesn’t mean shipping recklessly. There’s a line between “embarrassingly incomplete” and “dangerously broken,” and that line isn’t always obvious until you’ve crossed it.

We learned this the expensive way with Pathao Pay.

By 2017, we were growing fast. Rides were working. Food delivery was taking off. And we kept hitting the same friction point: cash. Everything in Bangladesh ran on cash. Customers paid cash. Riders collected cash. Merchants settled in cash. We were basically a logistics company running a parallel banking operation, and it was a nightmare.

So we decided to launch Pathao Pay. A closed-loop wallet. Customers could load money into the app, pay for rides and food digitally, and riders wouldn’t have to carry cash anymore. It was elegant. It solved real problems. And we built it fast.

Too fast.

We launched in beta. Started onboarding users. Got traction. Then, three weeks in, we got a letter from Bangladesh Bank. Not a friendly inquiry. A notice. We were operating what they considered a financial instrument without the proper license.

Technically, Pathao Pay was a closed-loop system. You couldn’t transfer money out. You couldn’t use it anywhere except Pathao. It was more like a gift card than a wallet. But we called it “Pay.” We marketed it like a payment solution. And in a country where mobile financial services are tightly regulated, that was enough to raise every red flag.

We had to shut it down. Immediately. Not “let’s phase this out over a few months.” Not “let’s fix the regulatory issues and relaunch in six weeks.” Immediately. As in, users woke up the next day and couldn’t access their wallets.

The fallout was brutal. Customers were furious. Some had loaded significant amounts. We had to manually process refunds. The press tore into us. Competitors whispered that we didn’t know what we were doing. And investors, the ones we were pitching for our next round, started asking harder questions about our due diligence processes.

It took us five years to relaunch Pathao Pay. Five years of working with regulators, applying for licenses, building compliance infrastructure, and proving we could operate within the rules. Five years of watching competitors move into the space while we sat on the sidelines.

The MVP logic that worked for courier, launch fast, learn from users, iterate in public, failed catastrophically for payments. Because the cost of getting it wrong wasn’t a bad user experience or churn. It was regulatory shutdown. And in fintech, you don’t get do-overs.

Spotify could launch with an incomplete music library. Airbnb could launch with founders taking photos themselves. We could launch courier with manual dispatch. But we couldn’t launch payments without proper licensing. The line between fixable and fatal isn’t about product quality. It’s about irreversible consequences.

What You Can’t Learn From a Deck

Here’s what most fundraising advice gets wrong. It focuses on the pitch. The deck. The narrative. The vision.

But when we raised our Series A, the pitch mattered less than what we could prove in real time.

Our lead investor didn’t just review our deck and write a check. They sent analysts to Dhaka. For weeks. These weren’t courtesy visits. They embedded with our operations team. They rode with our drivers. They sat in our call center listening to customer complaints. They watched our ops managers manually assign deliveries at 7 AM.

And every day, they asked the same question: “Can this actually scale?”

Not “does the deck say it can scale.” Can it. Right now. With this team. In this market.

We couldn’t fake that. We couldn’t polish our way through it. Either the system worked or it didn’t.

This is where the MVP becomes your pitch. When you’re raising a seed round, investors bet on potential. But when you’re raising growth capital, they bet on proof. And proof isn’t a forecast. It’s not a TAM slide. It’s watching your product work, in the wild, under stress, today.

We had bugs. Our app crashed sometimes. Our routing was inefficient. But we had something more valuable: live operations. We could show an investor our courier dashboard at 3 PM on a Tuesday and say, “Right now, we have 142 active deliveries happening across Dhaka. Here’s the average delivery time. Here’s the rider utilization rate. Here’s yesterday’s cash reconciliation.”

That’s what closed the round. Not the perfect product we’d build someday. The imperfect one running right now.

And when we had to explain the Pathao Pay shutdown? We didn’t spin it. We said: “We moved too fast. We underestimated regulatory risk. Here’s what we learned. Here’s the process we built to prevent it from happening again.”

Investors don’t expect perfection. They expect honesty and learning velocity. Show them you can ship fast, fail smart, and adjust quickly, and they’ll fund the next iteration.

The Dhaka Advantage

Silicon Valley loves clean solutions. Elegant APIs. Seamless user experiences. Frictionless everything.

Dhaka doesn’t care about any of that.

In Dhaka, the internet is slow. Smartphones are cheap and underpowered. Data is expensive. Addresses don’t exist in half the city. Traffic is unpredictable. And cash is still king.

So when we built Pathao, we couldn’t just copy Uber’s playbook. Our MVP had to work in the reality on the ground. Our app had to function offline. Our dispatch had to account for traffic that could triple without warning. Our payment system had to handle cash, because cards weren’t an option. Our address input had to use landmarks, because street numbers were fiction.

These weren’t “features.” They were survival.

But here’s the thing: these constraints made our MVP stronger. Every compromise we made for Dhaka’s reality became a competitive advantage. Uber’s app assumed reliable connectivity. Ours didn’t. When network coverage dropped in Mirpur, our riders kept working. Uber’s didn’t.

The MVP isn’t about building less. It’s about building for the real world, not the ideal one. And the real world, especially in emerging markets, is messy. The startups that win are the ones who embrace that mess early.

We launched ugly. We launched incomplete. But we launched functional. And in a city where nothing works perfectly, functional was enough to win.

I was standing in the middle of Gulshan 2, drenched in sweat, holding three packages that needed to be delivered across Dhaka before 6 PM. This wasn’t part of some grand strategy. This was desperation. We had promised Poter Bibi we’d deliver her products, and we didn’t have enough riders yet. So I became the rider.

The bus to Mohakhali was a tin can on wheels, packed so tight I had to hold the packages above my head to keep them from getting crushed. The smell hit me immediately - sweat, fried onions from someone’s lunch, diesel fumes seeping through broken windows. A man’s elbow dug into my ribs. The packages - wrapped in brown paper, addresses scribbled in marker - were pressed against my chest. My phone rang. Another merchant asking for pickups. I couldn’t reach my pocket to answer.

The distance between thinking about doing something and actually doing it is where most startups die. Not from lack of ideas. From the paralysis of overthinking.

The Pivot That Mattered

We launched Pathao as an instant delivery service. The whole pitch was built around speed. Get your package in an hour. Revolutionary, right?

Then Poter Bibi, a Facebook merchant selling sarees, gave us 20 deliveries and destroyed my worldview: “I don’t need instant. I need reliable two-day delivery but just reduce the proce.”

I could have argued. I could have waited to gather more data, run surveys, validate the new direction. Instead, I pivoted immediately. That same day, I started reaching out to other Facebook sellers. Within a week, we had signed up dozens who all wanted the same thing: reliable, affordable logistics. Not fancy tech. Just basics done right.

Why could I move that fast? Because the decision was reversible. If two-day delivery didn’t work, I could pivot again. We hadn’t signed ten-year contracts. We could test, learn, fail, and adjust.

When NOT to Move Fast

Here’s what nobody tells you: not all action is smart action. Some decisions can’t be undone easily. Those are the ones you need to slow down for.

In our second year, we were growing fast. I wanted to close funding quickly. A local investor offered to invest at very predatory terms, but I was desperate. I almost signed.

Fahim Saleh stopped me. “This is irreversible,” he said. “Once we take their money, we’re locked in. Let’s take two more weeks.”

I was annoyed. Two weeks felt like forever. But he was right. We slowed down, talked to more people. The valuation was not just low, the liquidation preferences were punishing, the investor wanted board control. If I had moved fast on that, we would have crippled the company.

Jeff Bezos talks about this as Type 1 versus Type 2 decisions. Type 1 decisions are one-way doors - hard to reverse, high stakes. Type 2 decisions are two-way doors - easy to reverse, low stakes. For Type 2 decisions, move fast. For Type 1 decisions, slow down and think.

The problem is we treat everything like Type 1. We overthink which email to send. We debate for weeks about trying a new marketing channel. But most decisions are reversible. Hiring someone on trial? Reversible. Testing new pricing? Reversible. Personally delivering packages? Reversible.

Signing away equity, locking into long contracts, choosing the wrong co-founder - those are irreversible. Think hard on those. Everything else? Just do it.

The Laziness Trap

The biggest enemy isn’t competition. It’s your own laziness disguised as “strategic thinking.” We tell ourselves we need more information. More validation. More certainty. But really, we’re just scared of doing uncomfortable work.

In Bangladesh, this manifests specifically. We love planning. We love meetings. We love PowerPoints with five-year roadmaps. But we hate the messy, unglamorous work of actually starting.

I saw this with our first tech hire. He spent three weeks designing the “perfect” dispatch system on paper. Beautiful diagrams. Elegant architecture. Zero code written. I finally told him, “I don’t need perfect. I need working. Build something ugly in two days that dispatches one delivery correctly. Then we’ll improve it.”

He looked offended. But that ugly system, held together with duct tape and prayers, is what we used to dispatch thousands of deliveries while he slowly built the better version.

Airbnb’s founders didn’t wait for perfect technology. When they noticed hosts uploading grainy photos, they grabbed a camera and went door to door in New York, personally photographing apartments. Bookings doubled. It wasn’t scalable. But it worked because they did it instead of talking about it.

Start Ugly

Your first version will be ugly. Your first attempt will fail in ways you can’t predict. You will feel stupid. But you have to start anyway.

When we first built dispatch software, it crashed constantly. Riders got lost. Packages went to wrong addresses. We had to call customers and apologize, then personally fix the mess. But each mistake taught us something. Each failure made the next version slightly better.

So here’s the framework: before you delay a decision, ask yourself one question - is this reversible? If yes, do it now. If no, take the time to think it through. Most things are reversible. Most actions can be undone, adjusted, or pivoted away from. The only thing you can’t reverse is time wasted overthinking.

I think about that bus ride sometimes. The smell of sweat and diesel. The packages pressed against my chest. The absurdity of it all. But that ridiculous afternoon led to conversations with merchants, which led to more deliveries, which led to understanding what customers actually wanted. None of that would have happened if I’d sat in a cafe waiting for a “better system.”

Don’t wait. Don’t overthink. If something needs to be done, korte hobe. Do it now. Ugly is better than nothing. And action, when it’s reversible, is always better than inaction.

The morning after we closed our first real funding, I realized something terrifying: I had to hire people. And I had no idea how.

I was 26. I’d never managed anyone. The only hiring I’d done was picking up riders off the street. But now we needed engineers. Real ones. People who would build something thousands depended on.

Here’s the math that kept me up: Newscred was hiring engineers in Dhaka for $3-4000. We could pay $300. Maybe $400 if we stretched. Plus equity that might be worth something. Or nothing.

We weren’t selling jobs. We were selling a bet.

The Corporate Transplant Who Couldn’t

Six months in, we thought we’d gotten smart. We hired someone from British American Tobacco, one of the biggest public companies in the country. He had the resume. Ten years of experience. Manager title. References that glowed.

He lasted three weeks.

The problem wasn’t his skills. It was everything else. At BAT, when he needed something, he sent an email to a department. At Pathao, if the printer broke, you either fixed it yourself or it stayed broken. When he hit a problem with ops, he wanted to schedule a meeting for next Tuesday. We needed an answer in the next hour.

“Where’s the marketing team?” he asked one day.

“You’re looking at him,” I said, pointing to myself.

He thought I was joking.

At big companies, people have lanes. You do your job, stay in your box, go home. At startups, if something needs to happen and you’re the closest person, it’s your problem now. This guy kept waiting for someone else to handle things. There was no one else.

They’re just conditioned for a different game. And you can’t de-program ten years of corporate conditioning in three weeks. We needed people who were already wired for chaos.

The Questions That Actually Mattered

In Bangladesh, hiring was its own category of impossible. In Silicon Valley, if you’re building something interesting, talented people will take a bet on you. In Dhaka, talented people want safety. Their parents want safety. Their in-laws have opinions. Society has expectations.

“My parents think startups are a scam,” one candidate told me during an interview.

“What do you think?” I asked.

“I think what you are doing is exciting - and I would rather do this than be a paper pusher”

I hired him.

You can’t screen for skills in a 45-minute interview. You can screen for something else, desperation in a good way. The hunger to build something that matters more than the fear of it failing.

I started asking different questions.

“Tell me about a time you failed and no one bailed you out.”

“If this company dies in six months, what do you want to have learned?”

“What’s the worst job you’ve ever had, and why did you stay?”

The answers told me everything. If someone talked about failure and blamed everyone else, they weren’t going to survive here. If they couldn’t articulate what they wanted to learn, they just wanted a paycheck. If they’d never stayed anywhere uncomfortable, they’d leave the second things got hard.

I also learned to test for adaptability. I’d give people a problem mid-interview. Something broken. A bug, a customer complaint, a logistics nightmare. I didn’t care if they solved it. I cared about whether they panicked or got curious.

The best hire we ever made was a kid who’d been freelancing, building websites for small businesses. His code was messy. His resume was two lines. But when I asked him how he’d handle a system going down during rush hour, he didn’t freeze. He asked three clarifying questions, sketched a plan on the whiteboard, admitted what he didn’t know, and said, “I’d figure it out.”

Within a year, he was leading a team.

Screening for Culture When Culture Is All You Have

In the beginning, culture isn’t something you write on a wall. It’s how you act when no one’s watching. It’s what you tolerate and what you don’t.

We didn’t have a handbook. But we had patterns.

If you screwed up and hid it, you were gone. If you screwed up and told everyone so we could fix it together, you stayed. If you needed someone to tell you what to do every hour, you didn’t fit. If you saw something broken and fixed it without asking permission, you were gold.

The cultural filter also came down to something uniquely Bangladeshi: how people talked about their families. In our culture, you don’t make career decisions alone. Your parents have a vote. Your spouse has a vote. Sometimes your uncle has a vote.

So during interviews, I started asking: “What did your family say when you told them you were applying here?”

If they lied and said everyone was supportive, I knew they hadn’t told their family. If they said their family was worried but they were doing it anyway, that told me something about their conviction. If they said they were still figuring out how to tell their parents, I respected the honesty, but I also knew they’d probably leave in three months when the pressure got too much.

The people who lasted weren’t the ones with the best degrees. They were the ones who’d already decided that approval mattered less than building something real.

Fighting Your Family to Join a Startup

The bigger battle wasn’t convincing candidates. It was convincing their parents.

In Bangladesh, there are exactly three acceptable career paths: doctor, engineer at a multinational, or banker. Startup founder? That’s what unemployed people call themselves.

I’d sit across from brilliant engineers who wanted to join us, and they’d say, “I need to talk to my family first.” A week later: “My dad says it’s too risky.” Or worse: “My mom is crying.”

This wasn’t unique to Bangladesh. But the intensity was different. Most of our potential hires were 24, 25, unmarried, living with their parents. Parental approval wasn’t a courtesy, it was a requirement.

I started inviting parents to the office. Showing them the space. Explaining the vision. Assuring them their son wasn’t joining a scam. It felt absurd. But it worked. Because once parents saw we were real, they relaxed.

The ones who joined anyway, against family pressure, became the core team. Because they’d already proven they could handle the hardest conversation they’d ever have. Everything else was easy by comparison.

Screening for the Invisible Resume

I stopped caring about credentials fast. Resumes told me where someone went to school. They didn’t tell me if they’d debug a server crash at 2 a.m.

I started asking three questions:

“Tell me about something you built that failed.” If someone couldn’t name a failure, they either hadn’t tried anything hard or couldn’t admit mistakes. Both were disqualifying.

“What’s the hardest argument you’ve had with a teammate?” This revealed whether they could disagree without being disagreeable. Whether they took feedback or got defensive.

“If we shut down in a year, what would you want to have learned?” This separated people chasing paychecks from people chasing growth.

I also watched how they treated our receptionist, our tea-maker, our junior staff. If you’re polite to me but dismissive to someone you think is “below” you, you don’t fit.

Netflix famously says they hire “fully formed adults.” We couldn’t afford that. We hired hungry, humble people who wanted to become something.

The People Who Stayed

Some of the best people we hired left. Better offers. More stability. And that’s fine. Not everyone is built for chaos.

But the ones who stayed weren’t there because we paid the most. They stayed because they felt ownership. When someone asked what they did, they didn’t say “I work at Pathao.” They said “I built Pathao.”

That’s the only hiring strategy that matters when you have nothing to offer: find people who want to build something, not just work somewhere. Give them a reason to care. Then get out of their way.

The founding team gets the credit. But the early team builds the company. The people who stayed when everything broke, when we couldn’t make payroll on time, when customers were screaming and competitors were poaching our riders, those people made Pathao real.

Skills can be taught. Belief can’t. And in the beginning, belief is the only currency that survives a crisis.

I used to think mentors were like cheat codes. Find the right person, ask the right questions, skip all the mistakes. That’s not how it works.

Fahim Saleh didn’t give me answers. He gave me access. When I showed him “IsTomorrowHartal?”, a joke website that went viral during the 2013 hartals, he didn’t say “great idea, here’s what to do next.” He said, “Come work with me.” Then he let me fail. Ten projects. Nine crashes. He watched me burn through ideas like matches, said nothing, let me learn.

That’s what good mentors do. They don’t save you from mistakes. They make sure you’re making mistakes worth learning from.

How to Actually Find These People

You don’t find your kitchen cabinet at networking events. You find them in the wreckage.

I met our first real advisor, the lawyer who saved us during the fundraising crisis, through desperation. We’d raised money but couldn’t access it for four months because Bangladesh had no clear understanding of how international capital can flow into Bangladeshi legal structure. I was calling everyone. Banks, law firms, random expats who might know someone. One call led to another. Someone said, “Try this guy, he did something with remittances once.”

That’s how it works. You don’t LinkedIn message successful people asking to “pick their brain over coffee.” You solve a problem publicly, and people who’ve survived that problem find you. Or you’re so deep in chaos that you have to find them.

When we were struggling with our first layoffs, I didn’t know any founders who’d done it before. So I asked our investors for intros. “Who in your portfolio has cut 50% of their team and survived?” Got three names. Two never responded. One got on a call, spent an hour walking me through the mechanics, the communication, the aftermath. He’s still in my cabinet.

The pattern: solve real problems, ask specific questions, follow referrals from people who’ve seen your work. Your kitchen cabinet builds itself through demonstrated need, not ambition.

What the Best Founders Actually Did

Jobs didn’t build Apple alone. He had Mike Markkula, an early Intel employee who’d retired at 32. Markkula didn’t just invest $250,000. He taught Jobs how to write a business plan, how to think about margins, how to present to investors. Jobs was a visionary. Markkula was a veteran. That combination built the foundation.

Zuckerberg’s kitchen cabinet was Sean Parker and Peter Thiel. Parker had built Napster, knew how viral products scaled and how to navigate Silicon Valley politics. Thiel had sold PayPal, understood venture dynamics and board management. Zuckerberg was 20. They’d already made millions and lost companies. That mismatch was the point.

Bezos surrounded himself with people from Wall Street, not tech. His kitchen cabinet included David Shaw, his former boss at D.E. Shaw. While other internet founders were chasing growth, Bezos was learning financial modeling and long-term capital allocation from people who’d managed billions. Different mentors, different company.

The pattern: founders pick kitchen cabinets that fill their blind spots. Jobs needed business discipline. Zuckerberg needed Silicon Valley fluency. Bezos needed financial sophistication. None of them picked mentors who looked like them. They picked mentors who’d survived what they were about to face.

When to Listen, When to Ignore

Here’s the framework I use: advisors get a vote, not a veto.

When we were deciding whether to expand to food delivery, our kitchen cabinet split. Half said focus on logistics, don’t dilute. Half said food delivery was the obvious next move, market was ready. Both sides had good arguments. Both had been right before.

I listened to everyone, then made the call myself. We expanded. It worked. But if it hadn’t, that would’ve been on me.

The decision matrix is simple:

Listen when: They’ve made your exact mistake in your exact market. A founder who’s raised in Bangladesh knows things a Silicon Valley founder doesn’t. Specificity matters.

Ignore when: Their experience is outdated or context-mismatched. Someone who built a company in 2005 doesn’t know how growth works in 2024. Someone who scaled in the US doesn’t know Dhaka’s regulatory maze.

Debate when: Two advisors you trust give opposite advice. That means the answer isn’t obvious. Get both perspectives, understand the reasoning, then decide.

The best advice I ever got came from someone who wasn’t even formally an advisor. Just a founder I met at a conference who’d sold his company. I told him about our growth. He said, “Growth without unit economics is just expensive dying.” It stung. It was also true.

Your kitchen cabinet should sting sometimes. If it doesn’t, it’s not working.

Because their job isn’t to make you feel good. It’s to show you what doesn’t work before it kills you.

There’s this myth that your business idea needs to be unique. That if someone else has already done it, then you’re too late, the door has closed, and you might as well give up. It’s a myth that kills more startups before they even begin.

Ideas are cheap. Execution is everything.

Pathao was not an original idea. We weren’t even the first motorcycle ride-sharing company in Bangladesh. There was a company called Share-A-Motorcycle that launched before us. They had a head start, the so-called “first-mover advantage.” But being first didn’t matter. What mattered was who executed better, who adapted faster, and who built trust with customers. On all three, we outperformed. And that’s why Pathao became a household name while Share-A-Motorcycle disappeared.

This pattern isn’t unique to us. Look at bKash, the crown jewel of Bangladesh’s fintech scene. Everyone thinks of it as a revolutionary idea, but it was modeled after M-Pesa, the mobile banking system that transformed Africa. The brilliance wasn’t in dreaming up something no one had ever seen before, it was in localizing the model, executing at scale, and building consumer trust in a Bangladeshi context.

Or take ShopUp, one of the most successful startups to come out of Dhaka. Their model is inspired by Udaan in India, a B2B platform for small retailers. ShopUp didn’t pretend to invent the category, they borrowed a proven playbook and adapted it to the needs of Bangladesh’s fragmented retail ecosystem. That’s what made them work.

The Idea Trap

The obsession with uniqueness is what I call the idea trap. Founders convince themselves that unless they stumble on something the world has never seen, they’re not entrepreneurs. They guard their “billion-dollar idea” like it is treasure, pitching it only in whispers, terrified someone will steal it. But here’s the uncomfortable truth: if your idea is so fragile that someone else can kill it just by hearing it, then it was never worth much in the first place.

Unique ideas are, in fact, more likely to fail. Because if nobody is doing it anywhere in the world, chances are the market doesn’t exist yet, or the timing isn’t right, or the problem just isn’t painful enough for people to pay for a solution. The graveyard of startups is littered with “unique” ideas that sounded good in theory but never crossed over into reality.

Meanwhile, founders who looked at what was working elsewhere and localized it have a much higher success rate. Rocket Internet built an empire copying successful Western startups and launching them in developing markets. Lazada was basically Amazon for Southeast Asia. Foodpanda was modeled on Grubhub. Ola in India was “Uber for India.” And they all became giants.

Execution Eats Originality for Breakfast

What makes or breaks a business is not whether you were the first to dream it up, but whether you can execute it better, faster, and smarter than anyone else.

When we launched Pathao rides, we weren’t pitching Dhaka something unheard of. Uber was already a global giant. Grab was dominating Southeast Asia. Gojek had already become indispensable in Indonesia. What we did was recognize that Dhaka’s roads, Dhaka’s traffic, and Dhaka’s people needed something local. We built for this city, not for Silicon Valley. That was our advantage.

Execution means obsessing over the small things. Recruiting drivers one by one when no one trusted the platform yet. Convincing skeptical students at BRAC University to take the first rides. Building manual Facebook groups before we even had a functioning app. And then scaling that chaos into something systematic. None of that required a unique idea, it required relentless follow-through.

Standing on the Shoulders of Giants

Some of the world’s greatest companies were not originals in the pure sense.

• Google wasn’t the first search engine, there was AltaVista, Lycos, Yahoo. They just built a better algorithm and executed relentlessly.

• Facebook wasn’t the first social network, Friendster and MySpace came before. Zuckerberg didn’t invent the category; he perfected it.

• Apple didn’t invent the MP3 player, the smartphone, or the tablet. They simply reimagined the user experience so well that competitors became irrelevant.

The pattern is clear: success belongs to the one who adapts and executes better, not necessarily the one who dreams first.

Localize, Don’t Idolize

The smarter path, especially in emerging markets, is to look at what works elsewhere and localize it ruthlessly. Dhaka is not Jakarta. Dhaka is not Silicon Valley. You cannot just copy-paste. You need to tweak the model until it bends to your reality.

Pathao’s early courier pivot is a perfect example. We thought people wanted instant delivery, like Postmates in the U.S. But the real need was reliability and cash-on-delivery for Facebook sellers. That was the pain point, not speed. Once we saw that, we shifted, and suddenly the model worked.

The Real Test

So the next time you catch yourself saying, “But someone’s already doing it,” stop. The real question isn’t whether the idea is unique, it’s whether you can execute it better, adapt it faster, and localize it smarter than anyone else.

Because at the end of the day, ideas don’t move markets. Execution does.