build a toy, not a company.

in the startup world, there's a concept called mvp—minimum viable product.

i like to call it the shittiest version of your product that gets the job done.

too many founders overcomplicate their first version. they think an mvp needs to be polished, feature-packed, and impressive. it doesn’t.

think of your dream product—the perfect version in your head.

now strip away 70% of it.

that’s your mvp.

your mvp isn’t about perfection. it’s about testing the core idea as fast as possible.

breaking it down: the uber example

let’s say you’re building the next uber.

you might think you need:

-a mobile app

-beautiful design

-real-time gps tracking

-messaging between driver and user

-notifications when the driver arrives

-rating and review system

wrong.

uber at its core is a way to request a ride and get picked up.

your mvp could be a simple web app.

no fancy design.

no ratings.

maybe not even real-time tracking.

just a way for a user to request a ride and for a driver to accept.

if that works, then you add features.

why simplicity wins

founders waste months—sometimes years—building complex features before launching.

by the time they release, they have no idea if users even want it.

features don’t define success. solving a real problem does.

focus on the core function first.

launch it.

get real users.

see if it works.

example: uber’s first version

uber’s first mvp looked nothing like what you see today.

it was clunky, basic, and far from perfect. but it worked.

same with zoom. same with doordash.

doordash’s 45-minute mvp

doordash’s founders didn’t build a full platform from day one.

their first version? a website called paloaltodelivery.com.

it had one page with a pdf of restaurant menus.

customers called them to place orders.

they used google docs to track deliveries.

then they delivered the food themselves.

built in 45 minutes. tested immediately.

only after proving demand did they scale.

paul graham’s rule: do things that don’t scale

early-stage startups shouldn’t automate everything.

instead, manually handle things to learn what works.

doordash founders delivered food themselves.

airbnb founders personally took photos of listings.

stripe founders set up payments manually for their first users.

this lets you validate demand before building complex systems.

how to build your mvp in 30 days

-launch fast → build the simplest version in 7-30 days.

-get 10+ users → test it with real people.

-collect feedback → ask what they love, what they hate.

-analyze carefully → don’t take every suggestion at face value.

-iterate → improve the product based on real user needs.

-repeat → keep refining through multiple cycles.

most founders skip the iteration step and assume they know what users want.

they spend 6+ months building, only to find out no one needs their product.

don’t do that. launch fast. learn fast. improve fast.