What is a Living ICP?
The ideal customer profile is often the most confident document in a firm and the least examined. It is written in a workshop, from aspiration, recent memory and the loudest win in the room. Then it is laminated into the sales deck, and from that day forward it aims the pipeline while nothing that happens in delivery is ever allowed to talk back to it. Sales pursues what the document says. Delivery discovers what the market says. The two accounts are rarely introduced.
A Living ICP closes that gap. A Living ICP is an ideal customer profile maintained as a working belief: built from the outcome record of which clients actually created value, updated as new outcomes are scored, and pruned segment by segment, as a recorded decision, when the evidence turns.
The workshop artefact problem
A profile written once decays in three ways. The market moves: the segment that was underserved when the workshop ran gets crowded. The firm moves: capabilities, pricing and the delivery model change what a good client looks like. And the profile becomes self-confirming: because the pipeline is filtered through it, the firm mostly collects evidence about the segments it already believed in, and almost none about anything else. The document does not merely go stale. It quietly steers the firm away from ever finding out that it is stale.
The cost is not abstract. Every campaign aimed by a decayed profile creates demand the firm will struggle to serve well, and every good-fit prospect outside the profile is demand the firm never creates at all.
What updates a Living ICP
Outcomes, not opinions. The raw material is the record of scored engagements: which clients renewed, expanded and referred; which engagements delivered the margin they promised; what it actually cost to win and to serve each one. That record accumulates in an Outcomes Ledger, and a living profile is answerable to it.
The evidence deserves weighing, not just counting. A renewal is measured evidence that a segment creates value. A projection of lifetime value is modelled. A partner’s conviction that “we always do well with insurers” is stated, and counts accordingly. Treating those three the same is how profiles rot politely. Weighting them by quality is the same discipline the evidence hierarchy applies to every other fact in the business.
Just as important is what should not update the profile: a single outcome. One spectacular win in a new segment is an anecdote, however vivid, and one painful loss is too. A belief that swings on the latest story is not being maintained; it is being startled. The profile should move when evidence accumulates, and hold still when it has not.
What should retire a segment
Segments leave a living profile the same way they enter it: by decision, against evidence. The signs are consistent and unglamorous. Engagements in the segment repeatedly cost more to win and serve than they return. Wins arrive but never renew. Delivery strains every time, for reasons that turn out to be structural rather than bad luck. When the record says a segment consumes more than it creates, the segment should be retired deliberately: the decision recorded, the evidence attached, the door left open for the facts to change.
The alternative is silent retirement, and it is worse than keeping a bad segment. The document still names the segment, so marketing still creates demand for it and sales still chases it, while everyone senior has privately stopped believing. The firm keeps paying to generate work it no longer intends to do well.
One concrete example
Clearly illustrative, with no customer implied. A services firm of a few hundred people wrote its profile three years ago around mid-market manufacturers. The outcome record since tells a different story. Manufacturing engagements won at a healthy rate, but renewals were rare and delivery strained on almost every one. Meanwhile a handful of logistics clients, never named in the profile, renewed, expanded and referred. A living profile reorders on that record: logistics rises on measured outcomes; manufacturing narrows to the one sub-segment where the outcomes were genuinely strong; the rest is retired as a recorded decision with the evidence attached. The next campaign is aimed by the record, not by the laminate. And when a demand signal arrives from a manufacturer next year, the firm can reopen the question with the reasoning still intact.
The profile is a belief, so maintain it like one
An ideal customer profile is one of the most consequential beliefs a firm holds: it decides where money, attention and capacity are pointed before any individual deal is discussed. Decision intelligence treats consequential beliefs with a consistent discipline: evidence attached with its quality labelled, updates driven by outcomes, changes recorded so they can be challenged. ONX applies exactly that discipline here: living ICPs are maintained from actual outcome data, from which clients created value, rather than written once in a workshop. The profile stops being a poster and becomes what it always should have been: the firm’s best current answer to a question it keeps asking.
Common questions
What is a Living ICP?
A Living ICP is an ideal customer profile maintained as a working belief rather than written once as a workshop artefact. It is built from the outcome record of which clients actually created value for both sides, it updates as new outcomes are scored, and segments are retired by recorded decision when the evidence turns against them. The profile stays answerable to what happened, not to what the firm hoped three years ago.
How is a Living ICP different from a traditional ICP?
A traditional ICP is written once, usually in a workshop, from aspiration and recent anecdote, and then rarely confronted with results. A Living ICP is the same profile treated as a belief under maintenance: every scored engagement outcome is evidence for or against it, the segments it names rise and fall on that evidence, and changes to it are deliberate, recorded decisions rather than silent drift.
What evidence should update an ICP?
Scored outcomes, weighted by their evidence quality. Renewals, expansions, delivered margin and referrals are measured evidence that a segment creates value. Kind words in a review meeting are stated evidence and count for less. A single spectacular win or loss is an anecdote and should move the profile very little on its own. Belief should move when evidence accumulates, not when a story is vivid.
When should a segment be retired from the ICP?
When the accumulated outcome record shows the segment consumes more than it returns: engagements repeatedly cost more to win and serve than they yield, wins rarely renew, delivery always strains. Retirement should be a recorded decision, made by a person against the evidence, with the reasoning kept so it can be revisited if the facts change. The failure mode is silent retirement, where nobody believes in the segment but the document still points the pipeline at it.