3 Reasons You Can't Find Your ICP

Hey y’all — here’s a puzzle for you:

Your startup’s product has highly engaged users who love it, but there’s no obvious common thread among them.

And without a clear ICP you’re not sure how to market the product, whose feature requests to prioritize, or what the longterm vision should actually be.

To make matters more confusing, each cut of the data you do also has a bunch of similar users but who are dramatically less engaged with the product.

A founder friend recently ran into this scenario and asked for advice. Here are a few ways I advised them to find your ICP in this type of scenario (and how you can too)…

3 Reasons You Can’t Find Your ICP

Ignore Personas

Most startups define their ICP based on the persona of their users.

Who are they? What are their job titles, or demographics? You create an avatar based on these, maybe even with a name.

You build the product for this persona specifically, asking yourself whether a proposed feature would be something this persona would use and love or not.

But there’s another way to group users, which is sometimes a better fit:

Their goals.

More specifically, defining your ICP by what your users are trying to use your product to accomplish rather than who they are.

The most well-known framework for this is the classic Jobs To Be Done framework from Clayton Christensen:

While JTBD has gotten some pushback in the last few years, most notably from former a16z GP Sriram Krishnan (see below), it can be useful in this specific scenario.

There’s precedence for this.

Notion became a unicorn despite not having a clear initial user persona.

Their early users weren’t the 25-40 year old knowledge workers who dominate their usage patterns today. Instead, they were a wide variety of people who were trying to replace messy personal wikis, spreadsheets, and docs with a single flexible system.

To find some signal here, look for shared context or triggering moments that are commonly mentioned in user interviews. Find similarities in quotes and go deeper with users to understand them more.

But here’s the counter argument for JTBD more broadly, as well:

You’re Earlier Than You Think

Sometimes the simplest explanation is the right one.

It might just be that you’re not missing an ICP, you’re just still in the discovery phase where multiple use cases are working but none is dominating (yet).

It could be that you’re not seeing how to cut the data the right way or, more simply, that the product isn’t mature enough yet.

A good way to get out of this phase is to look at specific high-growth adjacent data:

• Highest x-day retention

• Fastest time-to-value

• Highest referral rate

Focus on building the product to increase these things and see which users respond most positively. It’s potentially one of the few cases where adding “just one more feature” actually will get you closer to PMF.

The Market Velocity Isn’t Happening

Another option, however distasteful, is that the market you are trying to capture simply isn’t emerging quickly enough (or, in a worst case scenario, much at all).

This is the only truly bad scenario of the three because there’s no solution.

We know that Peter Thiel says to build a product for a small, fast-growing market but the key word there is fast-growing. Small is bad if it’s staying that way.

If the market is not going to reach venture-scale in a reasonable time-frame, you are best served to pack your bags and go back on the hunt for a new one.

Hopefully some of what you’ve built can be repurposed for it.

How We Can Help

Advertise in my newsletter to get in front of 60,000+ founders.