Generating Startup Ideas


In the seminal, ubiquitous and sometimes derided book Zero to One, the reader is sold on the idea that long-lasting, vertical progress should be invested in and executed.

This sounds great. I agree.

My copy: a parting gift from my last employers.
My copy: a parting gift from my last employers.

Vertical progress is to invent a new way of doing things; compared to horizontal progress, which is to either improve an existing system, or expand it.

Below is a simple process for generating your own zero to one idea. I hope you find it useful.

Before we begin, I’ll start with some background.

The Problem: Semantics

The phrases “vertical progress”, “zero to one”, “a 10x improvement” are bold and inspiring.

However, semantics can be daunting. A prospective founder could feel like they need an idea so brazen that it will take them from 0 to 1 in one step, before they start.

What they actually need is a problem which requires a 0 to 1 solution.

… And an idea that is 0 to 0.01, moving in the direction of that solution.

0 to 1 success is years of compounding effort, trial and error, success and failure, 7 days a week.

Just ask Facebook co-founder Dustin Moskowitz, when asked how it feels to be an overnight success:

“If by ‘overnight success’, you mean staying up and coding all night, every night for six years straight, then it felt quite tiring and stressful.”

Facebook started as a personal web page for Harvard students; before likes, posts and comments had been invented.

Apple started out with Steve Jobs selling 50 circuit boards to the local computer shop, funded by selling an old VW bus and a calculator.

AirBnB launched with a booking system which looked automated, but was actually a loader that spun until someone in the office frantically checked a spreadsheet before accepting the booking.

AirBnB's first "automated" booking system.
AirBnB's first "automated" booking system.

The list goes on.

So what do we learn from this? That the important thing is to get started.

Getting Started: The Old Way

An idea generation process I have seen play out previously goes something like: (1) generate an idea, (2) validate it, (3) decide if it’s good or whether to proceed.

A common pattern of brainstorming ideas and validating them at early stage.
A common pattern of brainstorming ideas and validating them at early stage.

Validating ideas at the concept level increases the likelihood of categorisation with existing companies, resulting in a false negative for the “is it vertical?” question. In turn, whitespaces in the market would be ignored, and therefore, so would vertical ideas.

Got a banking idea? Monzo’s done it. Social network? The Facebook empire has you covered. Back to the drawing board…

See more on that here.

With the old way, it’s tough to get started. Tough to mentally extricate oneself from the status quo.

What if there was a better option?

The New Way: Iterate Over Ideas to Build a Vision

What if you could start off with a list of all possible ideas for the problem you’re solving? Then you could simply run through each of them, kick the tyres a bit and end up with a shortlist.

Well - you can make the list yourself, quite easily.

For any given idea, simulate all possible elements you can dream of. An “element” being a product feature (but not a product itself); or a market feature (geography, user type, market position, distribution method, pricing, etc).

Then, iterate over different groupings of these elements. This will expand the previous process by at least an order of magnitude.

New process: second and third columns have been added.
New process: second and third columns have been added.

Note that validation is now applied to all possible element combinations per idea. This forces one to scan whitespace across the market, encouraging discovery of vertical ideas.

The Result: Spot the Wood for the Trees

Let’s review the impact of this process with a quick role play.

Imagine it’s 2011.

You’re an iOS developer and you think the taxi market is inefficient. The obvious thing to do is make an app for booking a taxi.

The Old Way

It’s human nature to design the app in your mind which replaces how the system currently works, rather than think of all ways of how it could work.

The current system is to call the private hire company, and tell them your journey details. Or, hail a cab and tell them the same information.

Your idea?

A web form you fill out with that information that is sent straight to a taxi company or driver. Much quicker!

Your first reaction might be that many private hire apps already exist, and in cities you can hail a cab anyway. If you made an app, private hire is the best place for supply of cars and they already have an app (UK’s largest private hire company launched their app in 2009), so won’t use yours. Plus it’s horizontal growth, not vertical.

So it’s a non-starter, right? Move on to the next idea.

👏 Well done! You may have just left billions of dollars on the table, for Uber, Lyft and incumbents to take.

What if you followed the process? Let’s see what happens.

The New Way

Step (1), think of a problem: taxis are inefficient. Think of a broad idea: a taxi app.

Step (2), think of all possible elements. App features include messaging, GPS, payments, user profiles, company profiles, etc. Market features include private hire taxis, freelance taxis (ie black cabs in the UK, yellow cabs in the US), etc.

Step (3), go through all groups and consider if this has created a good

💡 Remember: do not think of how the process currently works when generating ideas. Only think of this for discounting ideas after you have come up with them.

Let’s see where we get to:

Layer feature ideas on top of one another and think about how that impacts the market.  Rather than consider what is in the market and think about what is not there.
Layer feature ideas on top of one another and think about how that impacts the market. Rather than consider what is in the market and think about what is not there.

By getting to the end of the process, you realise the layered set of elements actually enables you to become (essentially) the largest private hire taxi company in the world, and define a new global market.

You would not have arrived at this conclusion had you not brainstormed the elements from first principles.

Bringing it All Together

Start with a problem in the market.

Then, do not think of a product that solves it, or consider what the status quo is in the market.

Ferociously brainstorm all possible elements (features of both product and market) that could work in this arena.

For all elements, take clusters of them together, and think about how that cluster would work as a product idea. Is it unique? Does it solve a need? Does it create a new market altoghether?

Best case: you invent a unicorn.

Likely case: you invent some awesome startup ideas that could be very successful.

Worst case: you don’t have any ideas, but you have a radically better understanding of the market you’ve thought about, and the products that already do (and don’t) exist. Maybe you have just increased the salary you can demand at work, the roles you can apply for, and the quality of content you can put online.

Good luck, and please do let me know how you get on. 😊

Onward! 🚀

📝 Subscribe

Enjoy this? You can get my latest articles, delivered straight to your inbox.