Skip to main content

Startup Ideas

·7 mins

Starting any company is hard. Startups are even harder. If you want to start a company but don’t have a business idea, then the first step is figuring out what to build.

Startup
Is this…a startup idea?

The main difference between a small business and a startup is your level of ambition, and how much money you can raise. Paul Graham defines it as “startup = growth”, but (most) startups only grow if they have resources (i.e., people, capital) to do so. The number of companies that start out bootstrapped, have growth, and never need to raise any money is extremely small, like infinitesimal. From a probabilistic perspective you’re better off trying to build things that are easy to fundraise for.

Consider Zapier—one of the rare truly bootstrapped successes. They reached $140M+ in ARR without VC funding, but for every Zapier, there are thousands of bootstrapped companies that never break $1M in revenue. The odds aren’t great.

Some Truths Before You Start #

Here’s a few things I’ve learned from living startups:

  • the founding team needs to be likable (or insanely talented), it will be extremely hard to raise money or hire talent unless people admire you in some way
  • at the end of the day, the only thing that matters is execution: you have to make things happen
  • fundraising is extremely hard unless you know people who can connect you with people who write big cheques (even better if you know them directly)
  • raising a lot of money doesn’t guarantee success: you need to have a product, something people (or companies) will pay for
  • great products can fail for a lot of reasons, but the #1 killer is probably lack of capital (if you have a great product you can buy your way out of a lot of problems)
  • the skills needed to fundraise are not the same skills needed for building great products or companies
  • you need to know when to quit, pivot, or move on entirely
  • bad founders kill companies at all stages, and in later stages you can’t fix it by raising more money

Just look at WeWork—they raised over $11 billion but still imploded spectacularly because of leadership issues. Meanwhile, Figma was built with just $67M (before their $20B Adobe acquisition) because they executed brilliantly on product and go-to-market1.

A Process for Generating Startup Ideas #

1. Get an Idea #

Coming up with startup ideas isn’t difficult, the hardest part is figuring out whether or not it’s worth pursuing. Start with a problem, something you experience every day, something you find frustrating or are passionate about, and something you want to fix. The best products are almost always things people built for themselves.

Stripe began when the Collison brothers were frustrated with how difficult it was to accept payments online. Airbnb started because the founders couldn’t afford their rent and had space for an air mattress. Notion came from the founder’s desire for a better note-taking tool. These personal pain points led to billion-dollar companies.

Startup ideas should meet a few minimum requirements, such as:

  1. It should be something you can build an MVP of by yourself. If you can’t iterate on the product it will be extremely hard to get things right.
  2. The market should be sufficiently large. The bigger the market, the bigger the opportunity. Look for problems that afflict nearly everyone everywhere.
  3. Is this a problem you can actually solve?
  4. Are you the right person to solve it?

One additional criterion that’s become increasingly important: does your idea have a clear path to sustainable unit economics? The era of growth-at-all-costs has faded since 2021. Today’s investors want to see a path to profitability.

2. Test It Out #

Testing out a startup idea is tricky, and sometimes good ideas test poorly because the market isn’t ready for the product yet (so maybe you should build it anyway). You should test for different things: customer appetite, and investor appetite. Try to find ideas that people will buy and invest in. If your product is ludicrously profitable and there’s insane customer demand, then perhaps you’ll never need to raise money, but that’s unlikely.

Here’s a few strategies you can use to test ideas:

  1. Build a simple MVP and get paying customers. The quality of the product doesn’t matter, the only thing that matters is if people are willing to buy it. Superhuman charged $30/month from day one and had a waiting list—that’s validation.
  2. Run an ad campaign with keywords that represent your product, and send people to an email signup landing page to gauge interest. If you can get a huge number of signups without spending much money, that’s a good sign people would be willing to pay for it.
  3. Look for investors, ask them for their feedback, and ask them specifically “is this something you would invest in?”. Again, the only thing that matters is whether or not someone is willing to part with their money to be a part of what you’re offering.
  4. Use no-code tools like Webflow, Bubble, or Airtable to create functional prototypes without writing code. This approach lets you test more ideas faster, and many successful startups began as no-code MVPs before building their own tech.

The founders of Instacart famously built their first version in just a few hours, ordered groceries through their own app, and delivered them personally. Within a month, they had enough traction to join Y Combinator, and within a few years, they were a multi-billion dollar company.

3. Get Some Buzz #

It’s always worth getting buzz if possible. If you can get media attention, or get to the front page of top of Hacker News or Product Hunt, it becomes much easier to start a company. Once your brand, product, or name is in the general consciousness of the startup community, everything gets easier.

Getting buzz without spending money is more of an art than a science. It’s about having the right timing, the right messaging, and offering something people want.

Unusual strategies can work wonders. Coinbase grew by giving away small amounts of Bitcoin to new users when crypto was still obscure. Loom used creative demo videos that went viral. Notion deliberately limited access early on, creating a sense of exclusivity that had people begging for invites.

Some buzz is also a great way to validate an idea. If you make it this far, all you need to do is execute.

Don’t Get Attached #

The worst thing you can do for yourself is getting attached to a bad idea, or a bad team. It’s important to know when it’s time to quit.

As part of the YC interview, they sometimes ask “are you going to continue working on this even if we don’t fund you?”. The ‘right’ answer is “of course I’m going to continue!”, but that’s not always the best thing to do.

If YC doesn’t want to fund you, that may be a signal that you need to think about doing something else, especially if everything else is not working. Keep in mind, however, YC does sometimes fund companies after multiple interviews. In those cases, the companies were usually already somewhat successful, with or without YC.

Stewart Butterfield (Slack) and Kevin Systrom (Instagram) both pivoted from failed gaming projects to billion-dollar communication platforms. Their willingness to abandon their original vision saved their companies.

Ideas Are Worthless #

Lastly, always remember that ideas themselves are worthless (unless you can sell an idea for an amount greater than $0, in which case I stand corrected). Even if you have a great unique special idea, there are probably 1 million other people with the exact same idea.

The only difference between success and failure is execution.

Just look at Facebook—it wasn’t the first social network (remember Friendster and MySpace?). Google wasn’t the first search engine. Uber wasn’t the first ride-sharing app. What set these companies apart was ruthless, relentless execution on a good-enough idea, not the originality of the concept itself.

Focus less on finding the perfect idea and more on becoming the kind of founder who can turn an average idea into an exceptional company.


  1. “Figma raises $50 million Series D.” TechCrunch, 2020. ↩︎