ICP analysis (ICP)
In short
ICP analysis is the research step that produces or updates an ideal customer profile. You examine closed-won, closed-lost, and churned accounts to see which company traits separate good customers from poor ones.
The output is a ranked set of traits and thresholds your team can turn into filters and scores. ICP analysis is the work; the ideal customer profile is the result.
What data to gather
- Won deals: Industry, size, region, tech stack, deal value, and sales cycle length.
- Lost deals: The same fields plus loss reason, so you can see where fit broke down.
- Retention and expansion: Which customers renewed, grew, or churned, and how long they stayed.
- Qualitative input: Notes from sales, success, and customer interviews on why they bought.
How to run an ICP analysis
1. Define success
Choose the outcome you care about, such as retained revenue after twelve months, not only closed-won. Otherwise you may select for fast but short-lived customers.
2. Compare groups
For each trait, compare how often it appears in your best customers against the rest. Say 40 percent of best customers are in one industry but only 5 percent of lost deals are. That trait is a strong candidate. These figures are an example only.
3. Look at combinations
One trait rarely explains much. Check pairs, such as industry plus size, since the combination is often where the pattern appears.
4. Validate on fresh data
Test the draft profile on accounts you did not use to build it. If it fails there, you found noise.
5. Operationalize
Convert findings to searchable filters and a scoring model, then share one definition across sales and marketing.
ICP analysis vs. ICP scoring
ICP analysis discovers which traits matter. ICP scoring applies those traits to rank individual accounts. Analysis usually happens a few times a year, and scoring runs continuously on your account list.
Common mistakes
- Small samples: A handful of deals can mislead. Treat thin data as a hypothesis.
- Survivorship bias: Studying only winners without lost deals hides traits that are common to both.
- Ignoring churn: A profile built on first purchase can target customers who leave.
- One-time exercise: Re-run it when products, pricing, or markets change.
ICP analysis FAQs
What is ICP analysis?
ICP analysis is the process of studying closed-won, closed-lost, and churned account data to learn which company traits predict good customers. The findings define or refine your ideal customer profile and feed targeting filters and scoring.
How do you do an ICP analysis?
Gather data on won, lost, and churned accounts, define what success means, compare trait frequency between your best customers and the rest, test trait combinations, then validate the result on accounts you did not use to build it.
What data do you need for ICP analysis?
You need firmographic and technographic fields for won, lost, and churned accounts, plus deal value, sales cycle length, retention, and expansion history. Interviews and sales notes add the reasons behind the numbers.
How often should you run an ICP analysis?
Run a full analysis at least once a year, and again after a major product, pricing, or market change. Check key results quarterly against new closed-won and closed-lost deals to catch drift early.
