What an ICP is for

An ICP is not a marketing artefact but a prioritisation tool. It answers one question: which
customers make the company grow fastest at the lowest cost?

Without a clear ICP, teams spread budget across an audience that either does not convert or leaves
quickly. Retailers without that focus routinely discover that 20% of customers generate 80% of
contribution margin — and that those 20% share a very specific profile.

ICP in B2B versus B2C e-commerce

In B2B an ICP is formalised as a set of firmographic traits:

Parameter Example for a SaaS vendor in e-commerce
Industry / vertical Fashion, DIY, electronics
Site traffic 500K+ monthly unique visitors
Annual revenue Above a defined threshold
Team An IT function plus marketing
Pain point No personalization, low conversion

In B2C an ICP is a behavioural profile of the shopper:

  • Purchase frequency: three or more purchases in 90 days, for instance
  • Order value: above the category median
  • Engagement depth: views five or more pages, interacts with recommendations
  • First purchase channel: organic versus paid, which carry different lifetime value

Building an ICP: from data to description

  1. Export the top 20% of customers by lifetime value over the last 12 months.
  2. Find the shared patterns: where they came from, what they buy, how often, what the first
    order looked like.
  3. Compare with the high-churn segment — the inverse patterns identify the customers to exclude
    from targeting.
  4. Write down three to five quantitative criteria — that is the ICP.

Tip: do not build an ICP from intuition or an internal survey. Use data. Subjective ideal
customers often differ from the ones that actually produce profit.

Using an ICP in personalization

Once formalised, an ICP moves into the CDP as a set of segmentation rules. From there it can drive:

  • Recommendation targeting: more assertive upsell strategies for the ICP segment
  • Content personalization: extended content in the categories those customers prefer
  • Lookalike audiences: scaling the ICP to similar new users in ad platforms
  • Product fit monitoring: a falling share of ICP customers among new signups signals a drift in
    the audience

Common mistakes

  • An ICP as an avatar rather than a segment: a narrative description with no measurable criteria
    cannot be operationalised in any platform
  • One ICP for every product: with a modular product, different modules attract different ICPs
  • A static ICP: markets move, and the customer who was ideal two years ago may no longer fit the
    current product