What is ICP fit?

ICP fit scoring assigns numerical weights to each attribute in your Ideal Customer Profile and calculates a composite score — typically on a 0–100 scale — for every account in your CRM or prospect list. The categories most teams score are firmographics (industry, company size, revenue, geography), technographics (existing CRM, marketing automation, complementary or competing tools), and strategic signals (recent funding, leadership changes, hiring activity in revenue functions).

Accounts are then bucketed into tiers: Tier A (80–100) matched on nearly every dimension and routed immediately to account executives; Tier B (50–79) with partial fit, placed into automated nurture with SDR follow-up; Tier C (0–49) deprioritized or excluded from outbound entirely.

The thresholds are calibrated against your own closed-won data. Pull your last 50–100 won deals, tag each with firmographic and technographic attributes, and you will almost always find that 70–80% of wins share three to five common traits. If your average won deal scores below 75 on your rubric, your cutoffs are too strict. If it scores above 90, the rubric is not differentiating well between good and great accounts — Tier A is too broad to be operationally useful.

Engagement signals — email opens, ad clicks, website visits — tell you that someone is paying attention right now. But attention from the wrong company is still the wrong company. A prospect who lacks budget authority, operates in a vertical where your product rarely lands, or sits below the deal-economics floor can exhaust a rep's calendar while never closing.

ICP fit is structural: it predicts whether a deal is possible before any sales motion begins. Research aggregated by Leadpipe shows that ICP-fit customers post net revenue retention of 110–130%, versus 70–90% for off-ICP customers — a gap that compounds every renewal cycle. Gainsight research similarly finds that companies with well-defined ICPs experience 24% lower churn and 33% higher expansion revenue.

Inflexion Point's analysis makes the sequencing explicit: ICP fit must be the first thing qualified, before MEDDPICC or SPICED, because no qualification framework can compensate for fundamental organizational misalignment. If the prospect fails the ICP test, the opportunity should never enter the forecast — and the discipline to walk away from bad-fit deals is a hallmark of top-performing sales organizations.

ICP fit and buyer intent answer different questions. Fit is about what a company is — its industry, size, tech stack, and structural characteristics. Intent is about what a company is doing right now — researching your category, visiting competitor pages, or downloading evaluation guides.

Best-practice scoring models keep them as two separate numbers rather than blending them into a single score. A high-fit, low-intent account is a good long-term nurture candidate — structurally attractive but not currently in-market. A high-intent, low-fit account probably should not receive heavy sales investment despite the apparent engagement signal.

The highest-priority tier — accounts that score well on both dimensions simultaneously — is typically only 5–10% of your addressable market at any given moment. That scarcity is exactly why filtering matters: it concentrates rep time on the small slice of accounts where timing and structural fit align.

ICP fit evaluates organizations; buyer personas describe individuals. The ICP answers 'should we sell to this company?' — it is an account-level filter applied before any contacts are reached. A buyer persona then answers 'who do we talk to inside this company, and how do we frame the conversation for them?'

In a B2B sale where Gartner estimates a typical buying committee consists of 6–10 decision-makers — each arriving with independently gathered information — both are necessary. A high-ICP-fit account with no mapped personas is an underexplored territory. A fully mapped buying group inside a low-fit account is effort misdirected.

The sequence matters: confirm account-level fit first, then identify the right personas within that account. Skipping the first step means investing persona research and multi-threaded outreach into accounts that can never yield an acceptable return.

A negative ICP — sometimes called an anti-ICP — is the explicit list of account types to exclude before scoring begins. Common disqualifiers include verticals that churn at 2x the average rate, companies below the deal-economics threshold where CAC exceeds recoverable LTV, technology stacks that are incompatible with your product, and organizational patterns like perpetual hiring freezes or procurement timelines that routinely exceed 12 months.

Practitioners consistently find that high-fit customers churn at significantly lower rates than low-fit customers — Leadpipe cites roughly one-third the rate. Building the negative ICP forces the historical analysis of churned and lost accounts that surfaces those patterns.

Without an explicit exclusion list, the same lessons get re-learned on new logos: a salesperson closes an account that looked close, it churns at month six, and the post-mortem reveals a pattern that appeared in three prior deals. The negative ICP converts that pattern into a pre-screen that prevents the fourth repetition.

ICP fit is the first filter Komo applies before any signal triggers a research or drafting action. When a buying signal fires — a champion changes jobs, a target account raises a funding round, a prospect starts hiring for a role your product supports — Komo checks the account against your ICP criteria before doing anything else. Only accounts that clear the fit bar enter the workflow.

This means the human reviewing a draft already knows the account is structurally qualified. The signal is timely, the fit is confirmed, and the draft is grounded in account research — so the rep's decision is whether to send, not whether the account is worth pursuing.

The combination — ICP-gated trigger detection, automated account research, and a human-approved send — produces the timing and relevance of a well-run signal-based motion without routing low-fit noise to the people who need to act on it.