👓 The PLGeek guide to ICPs in PLG

With template and examples

ICPs are essential in B2B, and nowhere more so than in PLG, because they form a shared language and understanding that spans and aligns even more cross-functional workflows and relationships than in sales-led companies.

This post is presented by Sprig - build a product people love.

Sprig is an AI-powered platform that enables you to collect relevant product experience insights from the right users, so you can make product decisions quickly and confidently. Next-gen product teams like Figma and Notion rely on Sprig to build the user-centric products people love.

At Snyk, Sprig was an indispensable part of our product stack, and now, readers of The Product-Led Geek can get 10% off!

Having a well-defined ICP is critical to

  • Engaging with the right segment to better understand what to build

  • Engaging with the right segment to better understand how to message what you’re building

  • Developing content with language-market fit (from paid to website to email to product and beyond)

  • Focus for new user acquisition efforts

  • Optimising onboarding for the right set of users & teams

  • Targeting for outbound sales

  • Targeting for account-based marketing (ABM)

  • Filtering and segmentation in the PQA scoring process

  • Running appropriate Product-Led Sales (PLS) playbooks

When you build a high-quality product for a specific segment of the market that shares a common set of problems, and you market around your ability to solve those problems in the language that the market segment relates to, and you acquire users and customers from that market segment, all of your PLG metrics will get a boost.

With this focus, you may initially acquire fewer users, teams and customers than casting a wider net, but you’ll activate, engage and retain them at much higher rates, fuelling future acquisition and engagement loops.

User-centric PLG teams will truly benefit from the focus that a well-defined ICP brings.

Serving non-ICP customers will distract and occupy precious time of product and other teams on non-strategic direction, and worse, lead product teams astray, building things that detract from your ability to serve your core market well.

Note: it’s really hard to resist the pull of potentially large dollar-value contracts that are outside of your ICP. I’m not here to advocate that you should always say no to those deals, but rather that you should be intentional with them. Try to understand the opportunity cost of the decision(s).

What should a good ICP definition include?

In my experience, the best ICPs balance enough detail to inform decisions and action, but provide enough latitude for discovery and experimentation. Overly detailed ICPs can bring too much complexity to S&M initiatives, making it hard to apply in real-world contexts. Simplicity can lead to more effective execution.

Additionally, markets and customer needs are always changing. If your ICP definition is too detailed, it can impart a subtle rigidity in how you operate and impact your ability to evolve along with changes in the market.

I recommend the following data points to be captured in your ICP definition, but if there is some important information for your specific context, go ahead and capture that too.

Market & Industry

Which market & industry do these customers operate in?

Example: Tech / Sales & Marketing


In which geographies do these customers operate?

Example: North America, EMEA, UK


How many employees do these customers have?

Example: 50-250 employees


What relevant problems do these customers face?

‘Relevant’ is important - these need to be problems that your product solves (or aspires to solve) well.

Example: Difficulty in managing client relationships, tracking sales activities


What are the implications of them not solving those problems?

Example: Lost sales opportunities, lower customer retention, reduced revenue

User roles

Who’s using the product? This is particularly important for horizontal products that can serve many use cases for many people - narrowing this down earlier on with the ambition to expand later can be beneficial.

Example: Sales Reps, Sales Managers, Customer Service Reps

Buyer roles

Who’s involved in the buying process?

Example: CROs, Head of Sales, Head of CS/CX

Buying process

How do they typically buy?

Example: Trial sign-up, self-serve evaluation, hand raise or triggered outreach, engagement with a sales representative, procurement approval


What is the typical technology landscape at these customers?

Example: Using spreadsheets or legacy outdated CRM systems, sales teams using digital communication tools (e.g. email, VoIP), and businesses emphasizing data analytics for customer insights.

Other relevant criteria

A catch-all for additional criteria that are not universal but might apply to your situation

Example: impacted by specific regulations/laws: GDPR for companies operating in EU, CCPA for those in California.

How to determine your ICP in PLG products

You might have a preconceived notion of your ICP.

Put it to one side for a moment.

Your ICP isn’t always who you believe you should be selling to.

It’s rare that PLG product companies have a well-defined, narrow ICP from day one before they get traction.

It’s usually a journey of discovery.

So, where should you start in defining your ICP?

1. Look at user & customer feedback

PLG doesn’t mean that product teams should have zero touch with customers. Quite the opposite, in fact. There’s simply no substitute for getting close to and talking with customers (and potential customers). The age-old advice of (increasingly metaphorically) getting out of the office is as true today as when it was first given.

Look for who gives the most enthusiastic glowing feedback (social media, product reviews, during demos, in support dialogues, moderated and unmoderated research, etc.). Continuously look for opportunities to talk with your customers.

It can also be very useful to look for negative feedback clusters - these are less likely to be your ICP.

Note: Just be careful with async feedback data where you don’t have opportunity to dig deeper - I’ve seen negative feedback patterns from ICP customers that turned out to be a combination of passionate and motivated customers, and a product that still had deficiencies.

2. Look at usage data

Of course, great PLG companies appreciate the importance of early instrumentation of products to uncover valuable behavioural signals. Look at this quant data to try to answer question such as:

  • Who is using the product most frequently?

  • Who is activating at higher rates?

  • Who is engaging with the product most deeply?

  • Who is retaining best over the long term?

Again, look at the negative patterns to exclude that type of company from your ICP.

  • Where is usage-based churn most prevalent?

  • Where is activation lowest?

3. Look at survey data

It’s a fallacy that you should strip every single point of friction from your onboarding process, including asking your new users to answer questions.

If the questions you are asking will not lead to valuable insight, or if you’re not going to leverage the data meaningfully (for example, in tailoring the user experience), then, by all means, strip those questions from your onboarding.

But make no mistake, asking pertinent questions during onboarding to help you learn more about new users and their teams and companies is a powerful lever.

It can even improve activation rates amongst your target audience because they become more psychologically invested in the product.

So don’t abuse it, but don’t shy away from it.

When it comes to providing useful input into your ICP definition, asking a few questions (during onboarding is recommended, but you can also run specific surveys amongst your existing user base) can provide a goldmine of data.

I recommend asking them about their role and the problems they’re here to solve. You can also ask about company size and industry (relying on enrichment may leave gaps in your data, at the very least when a user might initially sign up using a personal email address as they first explore your product).

Make all of these questions multiple choice, and offer an ‘other’ option as a fallback.

It’s also very useful to understand if they’re here for work purposes, personal, or education. So ask about that, too.

Users in your ICP will typically have no issue with taking the time to answer a few questions. Users who decide that answering a few questions is too burdensome lack motivation and, in all likelihood, are not your ICP.

Tip: Another fantastic survey mechanism to can provide useful input into ICP definition in PLG companies is running a Product-Market-Fit (PMF) survey and analysing results across different customer segments.

4. Look at marketing and sales outreach data

Another valuable source of data informing your ICP investigation comes from marketing and sales outreach. There’s a great mix of quant (ad campaign performance, email response rates, website conversion) and qual (feedback from cold outreach) data available here.

  • Who is converting better in paid campaigns?

  • Who is responding better to cold outreach?

  • Who is converting (e.g. to product signup) better from website landing pages?

Also, look at the negative patterns to exclude from your ICP.

  • Where is paid ad conversion the lowest?

  • Who is rarely responding to cold outreach?

  • Who has the lowest website-to-signup conversion?

5. Look at revenue data

With PLG, revenue is typically a lagging indicator of success, and operationally we’re focusing on more actionable metrics that lead to revenue gains. But revenue (particularly retained revenue) remains the ultimate high-confidence signal of meeting a market need well.

Look into your revenue data to answer questions such as:

  • Which segment has the highest dollar retention (lowest churn)?

  • Which segment is expanding most?

  • Where sales are involved, which deal cycles (of a comparable size) seem most frictionless?

  • Where is the time-to-convert shortest? (segment by self-serve and sales-led/assisted motions)

Once again, look at the negative patterns here, as those companies should likely be excluded from your ICP definition.

  • Where are you losing most often?

  • Where do you see most contraction and churn?

Note: In the losses, you may also find patterns that uncover a different ICP and opportunities for use case and product expansion, or even to build a new product.

6. Look at unit economics

Looking at unit economics data can reveal some very interesting insight in the context of your ICP definition.

  • Which segment has the highest LTV?

  • Which segment has the shortest CAC payback?

You shouldn’t determine your ICP exclusively on unit economics, but it can be an important additional lens to look through when narrowing your focus if yielding higher profitability is important.

For example, you may have identified a broader ICP definition that spans multiple verticals. It might seem logical since your product solves the same problem (or a similar enough problem) for them all.

However, it might be more expensive to acquire customers from some of those verticals than from others. Or you might find that some of those verticals (or other slices or segments) might exhibit higher churn rates than others.

Of course, all of this quantitative data should augment and be in support of the qualitative evidence you’re seeing and hearing through conversations with customers and prospective customers.

Note: while I’m not a fan of NPS for many reasons, as an additional input into researching ICPs it can be a useful indicator. Higher promotor scores will typically be associated with better fit customers.

Ultimately, remember that you want to find out the answer to these questions:

  1. Who is most acutely feeling the pain that your product solves?

  2. Are they willing to pay for that pain to be solved?

Expect your ICP to change

Last but not least, as hinted at earlier in this post, expect your ICP to evolve over time.

  • Users will start using your product in ways that you never envisaged

  • Your product strategy will evolve as you seek out growth opportunities

These things all have the potential to influence your ICP.

Acquisitions and organic multi-product expansion may even lead to you having more than one ICP.

Note: Having multiple ICPs is never a good place to start because a singular focus is paramount when trying to gain market traction.

As your ICP will naturally evolve, you need to periodically review and update your definition and re-align your teams so that everyone is always working against a definition that most accurately reflects your ideal customer.

Your ICP is not static, so consider your ICP definition as a dynamic artifact. 

Tip: Keep a changelog to note key changes in your ICP definition.

Tip: As you initially develop or evolve your ICP definition, it’s a great idea to test changes definition via easy-to-measure leading signals such as ad campaign performance, landing page conversions, etc. Feed what you learn back into continued iterations.

Template and example ICP

To close out this guide, here’s an example ICP definition for an API development and testing platform, along with a template.

Feel free to copy and use the template!

This post was presented by Sprig.

Until next time!

Join the conversation

or to participate.