Tony Sheng     Projects     Contact

How to force yourself to come up with startup ideas

For the first week of 2015, I came up with fifteen startup ideas every morning. Each one had to meet one simple criteria: related to two YC Request for Startups industries.

In October, I’d applied to YC with an app to help deliver mindfulness to people suffering from anxiety. It took everything I had to come up with and cultivate this idea, and my partner and I grew very attached to it. We were rejected. I didn’t have any ideas left.

I was planning to quit my job to do something more mission driven and aligned with my values. Something meaningful with the potential to improve the world in some way. Rejection threw a wrench in my plans.

Nonetheless, in December, I quit my job to begin a year of side projects (in hopes that one of them would catch fire and I could dedicate myself to it). While I’d saved up runway to support myself for a short period of time, I didn’t have enough ideas. So designing a repeatable process that helped me generate ideas became a first order problem for me to solve.

Ideation methods fall into two categories: organic and contrived.

Organic ideation arises from paying attention to the world around you and exposing yourself to lots of possible sources of inspiration. This is a sort of meditative practice of seeing the world clearly and observing carefully any thoughts that might turn into ideas and jotting them down immediately. I do this using Wunderlist and my phone. This is probably the best way for someone to come up with a good idea because it’s more likely to be related to something you know about simply because it arises organically in your everyday life. This process requires patience, though, so for our demanding purposes we’ll focus on contrived.

Contrived ideation means sitting down specifically to come up with ideas through various strategies. The strength of the strategy is measured by how effectively it produces ideas that fit your criteria. With startup ideas, my criteria was that each one had to relate to a YC RFS--this was the best heuristic I could think of that indicated a large opportunity.

One strategy could be to simply go down the list and try to jot down ideas for each of the industries, but with limited domain expertise in most of the industries, the ideas wouldn’t have much substance to them.

Instead, I tried to replicate the Medici effect. As author Frans Johansson says, “when you step into an intersection of fields, disciplines, or cultures, you can combine existing concepts into a large number of extraordinary new ideas.” It is what James Altucher calls “idea sex.” Everybody has thought of one idea in a given industry, but if you apply an insight from one industry to another, maybe a breakthrough occurs.

A strategy emerged: try to apply one industry to another and use that as a prompt.

ideasex screenshot

I built a simple script that took an array of industries and randomly generated prompts. It worked surprisingly well. Some of the prompts were hard, but I’m sure somebody else with more knowledge in that space might have an insight from it.

Now this doesn’t make the idea good because it doesn’t account for competition and whether or not you are positioned to execute. However, it does solve the problem of volume, which greases the gears to come up with ideas that are good.

“I’d do ___ if only I had an idea (or a technical cofounder lol)” is such a common excuse during YC application windows. Hopefully this little technique will help some people come up with more ideas this season.

Try some prompts on

If you enjoyed this post and want more like it, join thousands of readers on the weekly newsletter.