Your ICP (Ideal Customer Profile) is the specific type of customer who gets the most value from your product and is most likely to buy. Getting this right changes everything downstream, from prospecting to messaging to close rates.
Use this when: You're selling to everyone and closing nobody, or you've closed a few deals but can't articulate who your best customer is. You need a filter that turns "anyone who might buy" into "the specific person I should target."
You're done when: You can describe your ideal customer in one sentence — their role, company size, industry, and the specific problem that makes them a buyer — and use that description to build a prospect list in under 30 minutes.
The Sequence
1. Export your customer data
Pull every customer from your CRM or spreadsheet. If you're pre-revenue, pull every meaningful conversation from your discovery calls. You need at least 10-15 data points. For each, note: company name, contact name, title, company size, industry, deal size, how long they've been a customer, and how the relationship feels.
- Deliverable: A spreadsheet with all customers (or serious prospects) and their attributes.
- Common mistake: Skipping this step because you think you already know who your best customers are. Your gut is biased toward recent or large deals, not the full pattern.
2. Start from positioning, not demographics
Before you score customers, ask a more fundamental question: what are you the best in the world at, and who cares the most about that? Your ICP isn't a demographic exercise. It flows directly from your differentiated value. List the capabilities your product has that no competitor can match. Then ask: which type of customer values those capabilities so much that they'd switch from whatever they're using today? That's your ICP starting point. Demographics (company size, industry, role) are filters you apply after you've identified who needs your unique value the most. If you skip this step, you end up with a "target market" that looks reasonable on paper but doesn't actually convert, because it's not anchored to what makes you different.
- Deliverable: A short list of your differentiated capabilities and the customer characteristics that make those capabilities critical.
- Common mistake: Defining your ICP purely by firmographics without first identifying what you do differently. Two companies can be the same size, same industry, same stage — and one is a perfect fit while the other will churn in 90 days. The difference is whether your unique value matters to them.
3. Apply the four filters
Score every customer on four dimensions:
- Value: Which customers do you deliver the most value to? Where does your product make the biggest difference in their business?
- Revenue: Which customers pay you the most (or have the highest potential to)?
- Happiness: Which customers are happiest? Who renews without friction, refers others, and rarely complains?
- Enjoyment: Which customers do you actually enjoy working with?
That fourth filter matters. Customers you hate working with drain energy even if they pay well. They slow down your team, produce churn, and make your best people want to quit. Score each customer 1-5 on all four dimensions, then sort by total score.
- Deliverable: A ranked list of customers with scores across all four filters.
- Common mistake: Ignoring the enjoyment filter because it feels "soft." The customers at the intersection of all four filters are 3-5x more valuable over their lifetime than customers who score high on revenue alone.
4. Find the pattern
Look at your top 5-10 customers. Go to their LinkedIn profiles and note:
- Job title of your buyer
- Company size (employees or revenue)
- Industry or vertical
- Company age and growth stage
- Whether they're hiring (a proxy for growth and budget)
The pattern is your ICP. It's not a persona deck with stock photos. It's a search filter you can plug into LinkedIn Sales Navigator or Apollo and get a list of 500 prospects who look exactly like your best customers.
- Deliverable: An ICP statement: "[Role] at [company type] in [industry] with [size] who has [specific problem]."
- Common mistake: Making the ICP too broad. "Marketing managers at mid-size companies" isn't an ICP. "Head of demand gen at B2B SaaS companies with 50-200 employees who are hiring SDRs" is.
5. Validate the pain points
For your ICP, list the top 3 problems they face that your product solves. Use two sources: what your best customers actually told you during sales and onboarding, and a quick research loop. Ask ChatGPT: "You are a [ICP job title] at a [company type]. What are the 10 biggest problems you face?" Then iterate: "What solution to problem #3 would make you feel confident?" Cross-reference the AI output with what real customers said. Keep the problems that overlap.
- Deliverable: Three validated pain points, stated in your customer's language (not your product's language).
- Common mistake: Describing pain points in your terms instead of theirs. "Lack of workflow automation" is your framing. "I spend 10 hours a week copy-pasting between Salesforce and Slack" is theirs.
6. Concentrate your firepower
Before you build filters, understand why concentration matters. In most markets, a tiny fraction of accounts control most of the buying power. The goal of ICP definition isn't just to describe your buyer, it's to narrow ruthlessly so you can pursue fewer, higher-value accounts with more intensity. Once you have your ICP, use it to build a finite target list and pursue it with discipline: the same accounts, worked consistently for months, not a rotating buffet of new names every week.
- Deliverable: A prioritized target account list of 50-100 companies that match your ICP and represent outsized revenue potential.
- Common mistake: Rebuilding your prospect list every quarter instead of deepening your approach to the same high-value accounts. Consistency compounds. A prospect who ignored your first 5 touches may respond to the 8th.
7. Build your targeting criteria
Translate your ICP into specific filters you can use in prospecting tools:
- Job titles to target (and exclude)
- Company size range
- Industries
- Geography (if relevant)
- Trigger events (hiring, funding, product launches)
Test the filters by running them in LinkedIn Sales Navigator. If the results look like your best customers, the targeting is right. If not, adjust.
- Deliverable: A saved search or filter set in your prospecting tool that produces a list matching your ICP.
- Common mistake: Not spending enough time on targeting. It's boring work, but when you get it right, your reply rates, close rates, and customer quality all improve in lockstep.
Map the buying committee
Once you know who your ideal customer is, you need to understand who's actually involved in the purchase. In B2B, there's rarely one person making the decision alone. There's a cast of characters, and each one plays a different role.
Champion. The person inside the company who wants your product and will sell internally for you. They have the problem, they've felt the pain, and they'll push to get the deal done. Usually a manager or team lead. This is the person you spend most of your time with. If you don't have a champion, you don't have a deal.
Economic Buyer. The person who signs the check. They care about ROI, budget, and business impact. Often a VP or C-level. They might never use the product. They just need to believe it's worth the money. Sometimes the champion and the economic buyer are the same person, especially when you're selling to small companies or early-stage startups where the founder makes every purchase decision.
Gatekeeper. The person who can block the deal but rarely initiates it. CTOs, CIOs, security teams, legal, procurement. They look for reasons NOT to buy: compliance issues, integration risks, vendor policies. You don't sell to gatekeepers. You neutralize their objections. Figure out what they'll ask for (SOC 2, SSO, data residency) and have answers ready before they ask.
End Users. The people who will use the product daily. They care about UX, workflow fit, and whether it actually makes their job easier. Sometimes they're the champion too. Sometimes they have zero say in the purchase but their adoption determines whether the contract renews. Ignore them at your own risk.
In early-stage sales, one person often wears multiple hats. The champion is the buyer is the user. Your first 20 customers probably look like this. But as you go upmarket, these roles separate. A deal at a 500-person company might involve 4-6 people across 3 departments. Knowing who's who, and what each person cares about, is the difference between a deal that closes and one that dies in "internal review."
Template
Example
A startup selling API monitoring to engineering teams had 30 customers across fintech, e-commerce, healthcare, and SaaS. Running the four filters revealed a tight cluster: the top 8 customers were all platform engineering leads at fintech companies with 100-500 employees. They scored highest on value (API reliability is a compliance requirement in fintech), revenue ($1,200/month average vs. $400 for other segments), happiness (lowest churn rate), and enjoyment (technically sharp, fast decision-makers). The founder built a Sales Navigator search with those filters and got 340 results. That list became the entire Q2 outbound campaign and produced a 6% reply rate, triple the previous quarter's 2%.
Written with ❤️ by a human (still)