How To Conduct Ideal Customer Profile (ICP) Research That Actually Drives Results

If you've ever poured resources into a marketing campaign or sales push that went nowhere, there's a good chance the problem wasn't your messaging or your product. It was your targeting.

Knowing who you're selling to is just as important as knowing what you're selling. Yet it's surprising how many businesses, from scrappy startups to well-funded scale-ups, skip this step entirely or do it so broadly that the output is practically useless. The concept behind getting this right is called an ideal customer profile (ICP), and it's one of the most powerful tools in a data-driven marketer's toolkit.

In this article, I'll break down what an ICP is, why it matters, what a strong ICP profile actually contains, and how to conduct ICP research using both secondary (desktop) and primary (interview-based) methods. Whether you work in market research, product marketing, or analytics, this is a framework you'll keep coming back to.

What Is an Ideal Customer Profile?

An ideal customer profile (ICP) is a detailed description of the type of company or customer that is the best fit for your product or service. Think of it as a blueprint of your dream customer, the one most likely to buy, most likely to stick around, and most likely to get genuine value from what you offer.

ICPs are most commonly used in B2B (business-to-business) contexts, where the "customer" is an organisation, not an individual. But the concept applies equally well in B2C if you shift the focus from firmographics to demographics and psychographics.

The key word here is ideal. This isn't about describing every customer you could sell to. It's about defining the customers you should be pursuing — the ones where there's a natural fit between their needs and your capabilities.

💡 Pro-Tip: Don't confuse an ICP with a buyer persona. An ICP describes the company (or customer type) that's the best fit. A buyer persona describes the individual person within that company who makes or influences the purchasing decision. They're complementary, but they serve different purposes.

Why ICP Research Matters

If your ICP is too broad, your marketing will be generic, your sales team will waste time chasing bad-fit leads, and your churn rate will creep up because customers who aren't a good fit tend to leave.

Here's what a well-researched ICP actually does for you:

• Focuses your sales and marketing efforts. Instead of casting a wide net and hoping for the best, you target accounts and prospects that are genuinely likely to convert. This matters a lot when budgets are tight and teams are lean.

• Improves conversion rates. When you speak directly to the pain points, priorities, and language of a well-defined customer segment, your messaging resonates more deeply. People feel understood, and that drives action.

• Reduces customer churn. If you sell to the right customers from the start — customers whose needs align with what you actually deliver — they're far more likely to stay. Winning a bad-fit customer might feel like a win in the short term, but it often becomes a drain on your support and success teams.

• Aligns your teams. A clear ICP gives sales, marketing, product, and customer success a shared understanding of who you're building for. That alignment is powerful.

ICP vs. Buyer Persona: Getting the Distinction Right

This is a common point of confusion, so let's clear it up with a practical example.

Imagine you sell a cyber security services. Your ICP might be: a mid-market SaaS company with 200–500 employees, operating in financial services or healthcare, with a growing data team and compliance obligations that require robust data lineage and access controls.

Your buyer persona, on the other hand, might be: Sarah, a 38-year-old Head of Data who reports to the CTO, is under pressure to demonstrate compliance readiness, and evaluates tools based on integration with existing cloud infrastructure.

The ICP tells you which companies to go after. The buyer persona tells you how to talk to the people inside those companies. You need both, but the ICP comes first. Without it, you'll build personas that are detached from the reality of who actually buys.

What Does a Strong ICP Contain?

A robust ICP isn't a one-liner like "mid-sized tech companies." It's a multi-dimensional profile that covers several categories of information. The exact components will vary depending on your business, but here are the core elements most strong ICPs include.

Firmographic Attributes

These are the organisational characteristics of your ideal customer. Think of them as the "demographics" of a company:

• Industry or vertical (e.g. financial services, e-commerce, healthcare)

• Company size — typically measured by employee count or annual revenue

• Geography — where they're headquartered and where they operate

• Growth stage — startup, scale-up, enterprise

• Organisational structure — does the company have a dedicated data team, or is analytics handled ad hoc?

Technographic Attributes

What technology does your ideal customer already use? This is especially relevant for data and SaaS products:

• Tech stack (e.g. Snowflake, dbt, Tableau, Salesforce)

• Cloud infrastructure (AWS, GCP, Azure)

• Data maturity — do they have a modern data stack, or are they still running everything through spreadsheets?

Behavioural Attributes

How does your ideal customer behave in the market?

• Purchasing patterns — do they prefer self-serve or sales-led buying?

• Decision-making speed — is the sales cycle weeks or months?

• Brand loyalty — do they tend to stick with vendors or shop around frequently?

Pain Points and Needs

What problems keep your ideal customer up at night?

• Core challenges they face that your product or service addresses

• Triggers — what events or milestones push them to seek a solution? (e.g. a data breach, regulatory change, or rapid growth)

• Goals — what outcomes are they trying to achieve?

Psychographic Attributes

These are the values, attitudes, and priorities that influence how your ideal customer thinks and makes decisions:

• Risk tolerance — are they early adopters or do they wait for proven solutions?

• Values — do they prioritise speed, cost, innovation, or compliance?

• Culture — is the company data-driven or intuition-led?

💡 Pro-Tip: Be as specific as you can. A company with 101 employees and one with 699 are at completely different stages of maturity. "100–700 employees" is too broad to be useful. Narrow it down until your ICP feels almost uncomfortably specific — that's usually the right level.

How To Conduct ICP Research

Now for the practical bit. There are two main approaches to ICP research: secondary (desktop) research and primary research (such as interviews and surveys). Most effective ICP projects combine both, doing what we sometimes call research triangulation.

Approach 1: Secondary (Desktop) Research

Secondary research is about mining existing data and publicly available information to build a preliminary ICP. It's faster and cheaper than primary research, and it's where you should start.

Step 1 — Analyse your existing customer base

If you already have customers, your best ICP candidates are hiding in your own data. Pull a list of your top accounts, the ones with the highest lifetime value, the lowest churn, the shortest sales cycles, and the strongest satisfaction scores.

Look for patterns across firmographic and technographic attributes. Do your best customers tend to cluster in a particular industry? Are they a certain size? Do they use a specific technology?

If you use a CRM like Salesforce or HubSpot, you can run this analysis directly from your customer data. If you're working with a smaller dataset, even a well-structured spreadsheet will do the job. The goal is to surface commonalities among your most successful accounts.

Step 2 — Review win/loss data

Beyond your current customer base, examine deals you've won and deals you've lost. Are there patterns in the types of companies that tend to say yes versus the ones that walk away? Win/loss analysis is one of the most underutilised tools in B2B research, and it can reveal a lot about who your product resonates with.

Step 3 — Conduct competitive and market analysis

Look at who your competitors are targeting. Review their case studies, customer logos, and marketing messaging. This won't tell you everything, but it can help validate your hypotheses or reveal segments you hadn't considered.

Industry reports from sources like Gartner, Forrester, or the World Economic Forum can also provide valuable context on market trends, adoption curves, and emerging verticals.

Step 4 — Leverage digital intelligence tools

Tools like LinkedIn Sales Navigator, ZoomInfo, Crunchbase, and BuiltWith can help you enrich your ICP research with real data. For instance, BuiltWith can tell you what technologies a company uses, while Crunchbase can reveal funding stage and growth trajectory.

If you're using AI tools to assist with your research, tools like Claude, ChatGPT, or Perplexity can help you synthesise publicly available information, identify patterns, and even draft initial ICP hypotheses. Just remember to verify any AI-generated insights against primary sources before treating them as fact. And be sure to check out my other article on AI research tools, linked below:

🔗 8 Amazing Free AI Tools For Market Research

Approach 2: Primary Research (Interviews and Surveys)

Secondary research gets you a solid hypothesis. Primary research validates and deepens it. This is where you talk to actual customers, prospects, and internal stakeholders to pressure-test your ICP.

Customer interviews

Interviews with your best existing customers are the gold standard of ICP research. You're looking to understand not just what they bought, but why they bought it, when the need became urgent, and what alternatives they considered.

Here are some questions that tend to surface valuable ICP data:

• What was the trigger that made you start looking for a solution?

• What were you using before, and what wasn't working?

• How did you evaluate and compare options?

• What was the most important factor in your decision?

• Who else was involved in the buying process, and what were their priorities?

• If you had to describe the ideal vendor for this type of solution, what would they look like?

Aim for 8–12 interviews to start. That's usually enough to identify clear patterns without drowning in data. Record the interviews (with permission) so you can revisit them during analysis.

Lost prospect interviews

These are harder to arrange but incredibly valuable. Understanding why a prospect chose a competitor — or decided to do nothing — reveals the boundaries of your ICP. Sometimes the most useful insight is learning who you're not a fit for.

Internal stakeholder interviews

Your sales, customer success, and support teams interact with customers every day. They often have intuitive knowledge about which customer types are the best fit, even if they haven't formalised it. Run a short workshop or interview round with these teams to capture their insights.

Surveys

If you need to validate your ICP at scale, a short survey to your customer base can be highly effective. Focus on firmographic, technographic, and needs-based questions. Keep it to 10–15 questions to maximise completion rates.

💡 Pro-Tip: When conducting primary research, pay attention to the language your customers use. The specific words and phrases they use to describe their challenges and goals are marketing gold — use them in your positioning, website copy, and sales scripts.

Pulling It All Together: Building Your ICP Document

Once you've completed your research, synthesise everything into a single, shareable document. This doesn't need to be complicated, a well-structured one-pager is often more useful than a 40-slide deck.

Your ICP document should include:

1. A summary statement — one or two sentences that describe your ideal customer in plain language

2. Firmographic criteria — industry, size, geography, growth stage

3. Technographic criteria — tech stack, data maturity, infrastructure

4. Key pain points — the top 3–5 challenges your ideal customer faces

5. Buying triggers — the events or milestones that create urgency

6. Decision-making dynamics — who's involved, what they care about, how long the process takes

7. Disqualifiers — the attributes that signal a company is not a good fit (these are just as important as the positive criteria)

Share this document across your sales, marketing, product, and customer success teams. Make it a living document that you revisit and update regularly — at minimum, quarterly.

Common Mistakes To Avoid

Before I wrap up, let me flag a few pitfalls I see frequently in ICP research.

• Going too broad. If your ICP could apply to thousands of companies across multiple industries, it's not specific enough. Narrow it down until you can name 50–200 companies that genuinely fit the profile.

• Relying on assumptions instead of data. It's tempting to build your ICP based on gut feel or internal opinions. Don't skip the research. What your sales team thinks the ideal customer looks like and what the data actually shows can be very different things.

• Treating it as a one-time exercise. Markets change. Your product evolves. Your best customer today might not be your best customer in 12 months. Build regular ICP reviews into your planning cycle.

• Confusing ICP with total addressable market (TAM). Your TAM is everyone who could buy. Your ICP is the subset of that market where you have the strongest fit. Conflating the two leads to unfocused go-to-market strategies.

Conclusion

ICP research isn't glamorous, but it's foundational. When done well, it sharpens every downstream activity — from the messaging on your website to the accounts your sales team prioritises to the features your product team builds next.

The key takeaway is this: specificity wins. A narrow, well-researched ICP will outperform a broad, assumption-based one every time. Start with your existing customer data, validate with interviews, and build a document that your entire organisation can rally around.

If you found this article useful, subscribe to the Analythical newsletter for more actionable insights on market research, data analytics, and AI.