How to identify your ideal customer profile (ICP)

Part three of my seven-part series on kickstarting and scaling a B2B business

Aug 29, 2023

∙ Paid

👋 Hey, I’m Lenny and welcome to a 🔒 subscriber-only edition 🔒 of my weekly newsletter. Each week I tackle reader questions about building product, driving growth, and accelerating your career.

Welcome to part three of our series on how to kickstart and scale a B2B business:

Part 1: How to come up with a great B2B startup idea
Part 2: How to validate your idea
Part 3: How to identify your ICP ← This post
Part 4: How to find and win your first 10 customers
Part 5: How to find product-market fit
Part 6: How, and when, to hire your early team
Part 7: How to scale your growth engine

Let’s jump right in.

A huge thank-you to Akshay Kothari (COO of Notion), Ali Ghodsi (CEO of Databricks), Barry McCardel (CEO of Hex), Boris Jabes (CEO of Census), Calvin French-Owen (co-founder of Segment), Cameron Adams (co-founder and CPO of Canva), Christina Cacioppo (CEO of Vanta), David Hsu (CEO of Retool), Eilon Reshef (CPO of Gong), Eric Glyman (CEO of Ramp), Guy Podjarny (CEO of Snyk), Jori Lallo (co-founder of Linear), Julianna Lamb and Reed McGinley-Stempel (co-founders of Stytch), Keenan Rice (founding team), Mathilde Collin (CEO of Front), Rick Song (CEO of Persona), Rujul Zaparde and Lu Cheng (co-founders of Zip), Ryan Glasgow (CEO of Sprig), Shahed Khan (co-founder of Loom), Shishir Mehrotra (CEO of Coda), Sho Kuwamoto (VP of Product of Figma), Spenser Skates (co-founder and CEO of Amplitude), and Tomer London (co-founder and CPO of Gusto) for contributing to this series. Art by Natalie Harney.

We’ve so far spent all of our time finding and validating a problem, but surprisingly, many (perhaps most?) of the B2B companies I researched spent just as much time on picking the problem as they did on figuring out who to solve the problem for. When they didn’t do this, they often regretted it.

“We did not think about ICP. I wish we did earlier on. It’s one of my biggest mistakes.”
—Mathilde Collin, co-founder and CEO of Front

If you’ve got a killer idea but you’re talking to the wrong people, you’ll come away thinking your idea stinks, and give up. But the same idea pitched to different people can change everything.

Below, I’ll share a guide for nailing your ideal customer profile (ICP), dozens of stories of how founders identified their ICP, and as always, templates and tons of examples.

A few of my biggest takeaways and surprises from this step

Most founders initially got their ICP wrong.
Everyone landed on at least three attributes to describe their ICP.
Data from outbound sales is the best signal for what’s working, versus leads from investors and friends.
There are four common signs that you’re getting closer to your ICP:
1. A significant increase in your conversion rate
2. A significant increase in enthusiasm
3. A much stronger desire to take action now
4. The nod (see below)

Initial ICPs for some of today’s biggest B2B companies

Through my interviews, I’ve gathered the initial ICPs of over a dozen startups. In the chart, and in the stories below, notice how most companies landed on exactly three attributes. Some had more, but no one had fewer. Also notice how specific these attributes get.

How to identify your ICP

To identify your own ICP, go through the list below and take your best guess at picking the three most unique and important characteristics of your potential ideal customer (here’s a template to get you started):

Company size (e.g. 1,000-5,000 employees)
Job title (e.g. engineering manager, social media manager)
Pain point you’re solving (e.g. compliance, internal transparency)
Company’s unique way of working (e.g. design-driven, operationally heavy)
Specific tech used (e.g. data warehouse, GitLab)
Type of business (e.g. B2B SaaS, e-commerce)
Price point (e.g. sells software that’s $10k ARR)
Geo (e.g. urban centers, LatAm)
A unique place the user spends time (e.g. Node.js community)

Try to get super-specific and super-narrow with your ICP. Almost comically narrow.

Gusto found the most success with six ICP attributes initially:

“We started really, really, really, really, really, really, really narrow. Here’s how narrow:
Only companies that are five or fewer employees
in California
that offer no benefits
that have only salaried employees and no contractors
that have any other deductions
and that agree to get paid eight days after they run payroll
That ended up being a specific set of companies. The reason they chose us, and why we went after them, is because of the value that we provided.
Slowly, we expanded to where we felt ready. Let’s begin to serve hourly employees. Okay, now let’s help businesses provide benefits. Okay, now let’s move to additional states. We knew we were ready to expand when we felt like, one, we were getting customer love and, two, we had sufficient engineering bandwidth to go and build the features and products without letting go of the previous customer audience.
Today Gusto handles the bulk of what HR teams do, as well as some work of finance teams—payroll, insurance, benefits, onboarding, performance reviews, state registrations, tax credits, and more—for over 300,000 businesses with 1 to 500 employees in the U.S. and in 120 other countries. We’re 11 years in, but I still feel we are just getting started.”
—Tomer London, co-founder and CPO

The founders of Gong landed on three highly nuanced attributes:

“We tried to narrow it down as much as possible. We said we’re only going to sell to software companies that are:
Selling in the U.S. in English, so we could start with just one language.
Selling via video conferencing (it was Webex at the time), so we had the most amount of data to analyze.
Selling software that is worth somewhere between $1,000 to $100,000, because beyond $100k, we assumed it was going to be a different sales cycle, and less than $1,000, it would be too transactional.
There are probably 5,000 companies with this profile in the world. And exactly like the book Crossing the Chasm suggests, we thought, let’s get this small group working and get them excited. They would be the early adopters. We knew that we would understand their pain, and they’re typically advanced, so they would be early adopters.
And then of course, over time, we went wider, adding phone calls versus video conferencing, and then different company categories and geographies.”
—Eilon Reshef, co-founder and CPO

As did Snyk:

“Our initial target audience was a developer building with Node.js who was very security-conscious.
This depth-first approach was really important to validate the solution on the path to product-market fit. A JavaScript developer won’t care if you support Golang or Rust. Nailing the narrow and deep use case before expanding wider was critical.
The initial problem Snyk set out to solve was specific too: tracking and securing dependencies in the Node ecosystem. The community there would often discuss the inadequacies of the NPM dependency management capabilities. At the time, Node.js was gaining traction with increasing adoption in the enterprise, dedicated conferences, and the like, but it was still small enough that Snyk could meaningfully influence things. Here’s actually the first public showing of Snyk at Velocity Amsterdam in late 2015.”
—Ben Williams, former VP of Product

And Looker:

“Looker was a technical product, created right at the very start of data moving to the cloud and large, event style data being treated as business critical for analysis. So in 2013-2015, our core ICP was technical data teams—not end users or business analysts—who were starting to adopt cloud with large data sizes and complex analytical requirements AND the need to support a larger base on less technical end users within all functions of the company. These data teams tended to look at themselves more as engineering vs analysts, which aligned very well to our engineering first product design (code LookML, github built in, and all aspects of the product accessible via APIs). The companies tended to be startups with around 50-400 employees, with either engineering leading data or the data team being very technical.”
—Keenan Rice, founding team

Don’t stress if you can’t figure out your ICP for a while.

Here’s Ali Ghodsi, co-founder and CEO of Databricks (which is now worth over $30B) on their complete lack of an ICP early on:

“We didn’t have an ICP at all. In those years at Berkeley, we just wanted to change the world, honestly. We just wanted to have impact. That was the most important thing. We worked with all kinds of different companies. We worked with a hospital that was using this stuff, and they were doing genomic stuff. We worked with folks that were using us to determine earthquake magnitudes using Twitter. These are completely different customers and different user profiles. So, no, we didn’t have an ideal customer profile. We were just trying out whoever wanted to use it.”

And Rick Song, co-founder and CEO of Persona:

“We did the ICP exercise 17 times in the early days. So many times. Because at that time, all of the advice we ever got was like, ‘Know your ICP.’ So we kept trying to know who it was, and I also desperately did not want our ICP to just be startups. So we tried a lot. I think in earnest we never really had one, but we tried for sure.”

But the sooner you figure out your ICP, the faster things will fall into place.

“I didn’t think about ICPs whatsoever. However, looking back, that’s part of the reason PMF took so much time.” —Boris Jabes, co-founder and CEO of Census

“We did not think about ICP. I wish we did earlier on. It’s one of my biggest mistakes.” —Mathilde Collin, co-founder and CEO of Front

Eventually, you’ll be forced to figure it out.

“We didn’t focus on ICP at all, but once we did our public launch, we started getting a flood of different people coming in. Filtering through the leads spurred me into putting a much tighter definition around qualification for deals.” —Barry McCardel, co-founder and CEO of Hex

Start by paying attention to who gets most excited about what you’re building.

For Canva, here’s Cameron Adams, co-founder and CPO, sharing their path to identifying an ICP:

“We didn’t have any target persona until about six months into building the product. Through the conversations that we had with people, we began to see a certain segment really get excited about it. These social media managers, these people who were (a) figuring out what social media actually was, and then (b) having to scale this work across multiple clients and customers.
In 2012, Instagram had only really started. Pinterest had just come out. This whole notion of visual social media was still emerging. And Canva was, I think, the perfect tool at the right time for people who were grappling with what visual social media was. The social media manager/blogger audience was definitely the perfect one for us to start building a community around—especially freelancers, who were building their own social media management business. We gave them early demos of the product and then helped shape the feature set through their feedback. They were probably the ideal launch customer.”

For Sprig, Ryan Glasgow, founder and CEO, on narrowing in on their unusual initial segment:

“We found pretty early that our ICP was companies with millions of users. Being an early-stage startup selling to larger companies was actually one of our biggest early challenges. You have product-market fit for companies that are more difficult to sell to.
Our first customer, Thunkable, had millions of users. Square and Robinhood—also two of our early customers—had millions of users. I was getting very strong pull with companies like these. The larger the user base, the more excited they were about Sprig. Robinhood agreed to a large contract even though we were early. They’re like, ‘We love it. We’ll install tomorrow.’ They installed it as we were building the first version, basically. So that told us to go after these larger companies, with millions of users.”