Most ideal customer profiles are written in a workshop. They describe the customer a team wishes it had, not the one it actually wins. The fastest way to a useful ICP is to look at the deals you have already closed.
This is the method we use at the start of every B2B lead generation project.
ICP vs buyer persona
An ideal customer profile describes the companies you should sell to. A buyer persona describes the people inside those companies who influence or make the decision. You need both: the ICP tells you which accounts to target, and personas tell you who to contact and what to say.
Start with the ICP. A great message sent to the wrong company still fails.
Step 1: Pull your closed-won deals
Export the deals you have won over the last one to two years from your CRM. Include deal value, sales cycle length, source, industry, company size, location and the role of your main contact. If you have enough data, include lost deals too for comparison.
Step 2: Rank customers by value, not just count
Not every customer is a good customer. Rank them by revenue, retention, expansion and how easy they were to win and serve. Your ICP should describe the top of this list, not the average.
Include customers who were easy to onboard and stayed engaged, not only the largest deals. A customer who renews, expands and refers others is often worth more over time than a single large contract that churns after a year.
Step 3: Look for shared traits
Compare your best customers and look for patterns:
- Firmographics: industry, company size, revenue band, location
- Technographics: tools they use, such as a specific CRM or platform
- Situation: growth stage, team structure, recent changes
- People: the role that found you, the role that signed, who else was involved
Pay attention to what is missing as well. If you never win deals below a certain company size, that is part of your ICP too.
Segment before you generalise
If your best customers fall into clearly different groups, such as two industries or two company sizes, analyse each group separately. Averaging them produces an ICP that describes nobody. Many companies end up with one primary ICP and one or two secondary segments, each with its own criteria and messaging.
Look at sales cycle length and deal size per segment as well. A segment that closes quickly with smaller deals may suit outbound and paid campaigns, while a slower segment with larger deals may justify account-based marketing.
Step 4: Identify buying triggers
A trigger is an event that makes an account more likely to need you now. Look back at your best deals and ask what had just changed for each customer. Common triggers include hiring for a specific role, new funding, leadership changes, expansion into new markets or adopting a new tool.
Triggers turn a static list into a timely one. They are also what make outreach relevant, because the message can reference something real.
Learn from lost deals and churned customers
Closed-won data tells you what success looks like. Lost deals and churned customers tell you what to avoid. Compare them with your best customers and look for warning signs, such as a company size you rarely win, an industry with long cycles and low close rates, or a missing role that usually means no budget.
Adding exclusion criteria to your ICP saves outbound time and keeps sales focused on accounts that can actually close.
Example ICP scorecard
A scorecard turns findings into something the whole team can apply. For example:
- Industry: B2B software or IT services (high weight)
- Company size: 50 to 500 employees (high weight)
- Tech stack: uses a modern CRM such as HubSpot or Salesforce (medium weight)
- Trigger: hiring for sales development or marketing roles (medium weight)
- Region: a market you already serve (low weight)
- Exclusions: existing customers, competitors, companies below a minimum size
Each account gets a score, and the score decides whether it goes into active outbound, ABM, nurture or nowhere.
Step 5: Write the ICP as a scorecard
Turn the findings into a short scorecard your whole team can use. Each criterion gets a weight, and every account gets a fit score. This is the fit half of a lead scoring model, and it gives marketing and sales one shared definition of a good account.
Interview your best customers
CRM data shows what your best customers look like. Interviews show why they bought. Speak to five to ten of your best customers and ask:
- What was happening in the business when you started looking for a solution?
- What did you try before, and why did it not work?
- Who else was involved in the decision, and what did they care about?
- What nearly stopped you from buying?
- What would you tell a peer considering us?
Their answers reveal triggers, objections and the language buyers actually use. Record the exact phrases, because they are often better than anything your team would write for messaging, ads and content.
Step 6: Build and enrich the account list
Use data providers such as Apollo, ZoomInfo, Clay and LinkedIn Sales Navigator to find companies that match the scorecard, then enrich them with decision-makers and trigger data. Verify every contact before outreach to protect cold email deliverability.
Store the fit score and tier on the account record in your CRM, not in a separate spreadsheet. When the ICP lives where your team works every day, it actually shapes who gets contacted, which campaigns run and how deals are prioritised.
Step 7: Use one list across every channel
The same account list should drive outbound, account-based marketing and paid audiences. When every channel focuses on the same accounts, each touch reinforces the others.
Write personas for each target role
Once accounts are clear, describe the two or three roles that matter most in each deal. For each, capture what they are responsible for, which problems they feel most, how they measure success, and what objections they usually raise. These notes drive messaging for cold email, LinkedIn outreach and content.
Common ICP mistakes
- Defining the ICP from ambition instead of data
- Making it so broad that it excludes almost nobody
- Ignoring triggers, so the list is right but the timing is wrong
- Keeping the ICP in a slide deck instead of the CRM
- Never updating it as the product and market change
Turning the ICP into messaging
An ICP is only useful if it changes what you say. For each segment, write down three things: the problem they feel most, the outcome they care about, and the proof that makes you credible to them. That becomes the core of your outreach, ads and landing pages.
For example, a 50-person SaaS company hiring its first SDRs worries about wasting money on a channel that has not been proven. A 400-person IT services firm with an established team worries about pipeline consistency. Same category, very different messages.
Using triggers in practice
Triggers work best as a reason to reach out, not as a gimmick. A message that says "noticed you are hiring three SDRs, so outbound is probably becoming a priority" connects a real event to a real problem. A message that only says "congratulations on the new role" does not.
- Choose two or three triggers that genuinely predict a need for what you sell
- Set up alerts or regular searches so new trigger events reach your team quickly
- Reach out while the trigger is recent, usually within a few weeks
- Write the message around the problem the trigger suggests, not the trigger itself
Tiering accounts for different channels
Not every account deserves the same effort. Tier your list by fit score and deal potential. Top-tier accounts get coordinated account-based marketing with tailored content and senior outreach. Middle-tier accounts get personalised sequences. Lower tiers get lighter-touch campaigns and nurture. Tiering lets a small team focus its time where it matters most.
Review it every quarter
Your ICP will change as your product and market change. Revisit it each quarter with fresh closed-won data, and update the scorecard when the pattern shifts.