start specific. abstract later

Every podcast lies to you about the first two years. Zerodha’s founders talk about philosophy now. Go look at year one. It was firefighting. Keep the lights on, keep the customer happy, make sure money came in. Nobody talks about year one because it’s not worth talking about. That’s exactly why you should study it.

Founders who start from the vision almost never get to specificity. The factory owner doesn’t know he has a coordination problem. In his head the system is as efficient as any he’s seen. Tell him you’re solving fragmentation and he’ll smile, say nice things, and not pay you. Tell him his accountant fills the e-way bill by hand on Friday nights and you can automate it, and he pays.

Airbnb didn’t launch an app. They found one conference in San Francisco, five houses, and got the organisers to put it in a newsletter. Hotels in San Francisco didn’t serve breakfast, so the site was airbedandbreakfast.com. Then they noticed hosts were too lazy to take photos, so they sent two photographers per city. Flipkart’s founders stood outside a Bangalore bookstore and said go in, pick a book, I’ll deliver it cheaper.

Uncool, every time. That’s the pattern.

So the move is: smallest possible problem that someone pays for. Not a pilot, not an MOU. Pay. If they’re not paying it’s not a real bottleneck. Then go to their competitor and say I did it for them. After twenty of them you can start combining and calling it a platform. That’s two years out, minimum.

One hack: don’t go to family. Anyone who knows you will take you for granted. Go to a similar business where a 30 year old just took over from their father. They’re easier to convince than the 60 year old, and once you’ve done it for them, the whole industry can see it.

And keep two stories. The vision is for VCs. The local step is for the customer. You don’t mix them. A founder’s whole life is the gap between the step that looks too small and the vision that looks too big. The only way through the gap is to make the small step work so well that the big one stops looking crazy.

Don’t even build yet. Make a landing page, show it to them, iterate until someone says I’ll pay 10k a month. They don’t know there’s no backend. That’s fine.

the call where i said it best

She wanted to fix fragmentation in manufacturing. I told her to read The Goal by Eliyahu M. Goldratt over a weekend. Then go to one factory, talk to the workers not the managers, and map the pipeline on a Miro board. Raw material arrives, two hours to unload, sits two weeks, machine one, machine two. Somewhere there’s one machine that’s the bottleneck. Double it and throughput doubles. Tell the owner you can 2x his output and ask if he’ll pay. He will.

do this this week

Take the biggest problem you want to solve and find the smallest paid version of it inside one customer. Go sell that.

the line to remember

smallest possible problem that you can solve, which people pay for. if they are not paying, they are just taking you for a ride.