What Is ICP Match Rate?

ICP match rate is the percentage of your identified website visitors who actually fit your ideal customer profile. It is a different number than your identification rate (how many visitors you can name) and a more important one, because a high identification rate full of the wrong companies does nothing for pipeline. Most GTM teams report the wrong metric to leadership: they celebrate "we identified 60% of our traffic" when the number that predicts revenue is "38% of what we identified is actually who we sell to."

If you've never calculated this number for your own site, you're not alone, and you're also not optimizing the thing that actually matters.

Identification Rate vs. ICP Match Rate: Why the Difference Matters

Identification rate answers one question: can we put a name on this visitor? Person-level identification typically resolves 40 to 65% of US B2B traffic today, and account-level identification typically resolves 80 to 95%, depending on the vendor and the traffic profile (Knock2's own published benchmarks are 93%* account-level and 62%* person-level, measured against engaged sessions, defined as any visit of 10+ seconds or 2+ pageviews).

ICP match rate answers a completely different question: of the visitors we identified, how many are companies and people we'd actually want as customers? A visitor identification tool with a mediocre identification rate but a strong filtering layer can outperform a tool with a best-in-class identification rate and no ICP logic, because the second tool just hands your SDRs a longer list of the wrong accounts.

This is the gap that AI agent traffic and false-positive identity matches both exploit. Both inflate your identification rate while quietly dragging your ICP match rate down, and if you're only watching the first number, you won't notice until your SDRs start complaining that the "hot leads" from your visitor ID tool never convert.

How to Calculate Your ICP Match Rate

The formula is simple. The discipline to run it consistently is where most teams fall short.

• Step 1: Pull every uniquely identified company (or person) from a fixed window, 30 days is a good starting point.
• Step 2: Score each one against your written ICP definition (see the weighting framework below). Don't eyeball it; every visitor gets a score.
• Step 3: Set a pass threshold. A common starting point is requiring a "High" or "Very High" fit rating on at least two of your four weighted signal categories.
• Step 4: Divide the number that pass by the total number identified. That's your ICP match rate.

Run this monthly, not once. ICP match rate drifts as your traffic mix shifts, as you launch new campaigns that attract different visitor types, and as bot and agent traffic patterns change. A number you calculated in Q1 tells you nothing about Q3.

The ICP Fit Scoring Framework

Most teams that try to build an ICP score end up with a vague, unweighted checklist. The fix is to weight signals by how strongly they actually predict a closed-won deal for your business, not by how easy they are to collect. A practical starting framework, ranked by typical predictive strength for most B2B teams:

• 🔴 Firmographic fit - industry, employee count, and revenue band matched against your closed-won customer base. This is usually the single strongest predictor and the one teams weight the least.
• 🟠 Technographic fit - whether the identified company runs the tools your product integrates with or replaces. Especially predictive for infrastructure and workflow tools.
• 🟡 Engagement depth - pages visited, time on site, and whether high-intent pages (pricing, integrations, comparison pages) were part of the session. Real signal, but easy to fake with bot traffic, so pair it with traffic-quality filtering.
• 🟢 Geographic fit - whether the visitor is in a region your sales motion actually covers. Low predictive power on its own, but a fast disqualifier.
• ⚪ Company growth signals - hiring velocity, funding stage, tech stack expansion. Useful as a tiebreaker, rarely decisive alone.

Assign each category a weight that reflects your own closed-won data (RevOps should own this, not marketing), score every identified visitor against it, and route only the visitors that clear your threshold into active outbound, whether that's a lead scoring model feeding your SDR queue or a direct automated workflow to Slack and your CRM.

What Drags Your ICP Match Rate Down

When teams calculate this number for the first time, it's almost always lower than they expected. Three causes show up repeatedly:

• Automated and agent traffic inflating your denominator. If AI agents, crawlers, and monitoring bots are getting scored as identified visitors, they're diluting your match rate even though they were never going to become customers. See our breakdown of how agent traffic corrupts visitor ID data for a filtering framework.
• Identity false positives. A wrong-company or wrong-person match doesn't just waste a record, it actively pollutes your ICP scoring because you're evaluating fit against the wrong entity entirely.
• An ICP definition that's too broad to be useful. "Series A to Series C SaaS companies" isn't an ICP, it's a market. If half your traffic passes your filter, the filter isn't doing its job.

What's a Good ICP Match Rate?

There's no universal number, your ICP match rate depends on how narrow your ICP is and how targeted your traffic sources are. But as a rough industry orientation: companies with a tightly scoped ICP and demand gen that's actually targeted at that ICP typically see match rates in the 15 to 30% range of total identified traffic. Broad-market products with wide top-of-funnel content commonly see single digits, which isn't necessarily a problem if the absolute volume of ICP-fit accounts is still healthy. The number to watch isn't the percentage in isolation, it's the percentage trending over time as you tighten targeting and traffic quality.

Turning ICP Match Rate Into a Working Habit

Knowing the number is step one. The teams that actually benefit from it build it into how they route and prioritize: identified visitors above the ICP threshold get fast-tracked into person and account-level identification workflows and SDR outreach, visitors below it get nurtured or ignored, and the whole calculation gets re-run monthly against fresh closed-won data so the weighting keeps improving. That feedback loop, more than any single tool, is what turns "we get a lot of website traffic" into a pipeline number your CFO cares about.

Frequently Asked Questions

What is ICP match rate in B2B sales?

ICP match rate is the percentage of identified website visitors (people or companies) that meet your defined ideal customer profile criteria, calculated by scoring each identified visitor against weighted firmographic, technographic, engagement, and geographic signals and dividing the number that pass by the total identified.

Is ICP match rate the same as identification rate?

No. Identification rate measures how many visitors you can put a name or company to. ICP match rate measures how many of those identified visitors are actually a fit for what you sell. A tool can have a high identification rate and a low ICP match rate at the same time.

How often should I recalculate ICP match rate?

Monthly at minimum. Traffic mix, campaign targeting, and bot/agent traffic patterns all shift faster than most teams expect, and a stale ICP match rate calculation will misdirect SDR time.

What lowers ICP match rate the most?

Automated and AI agent traffic being counted as identified visitors, identity false positives (wrong company or wrong person matches), and an ICP definition that's too broad to meaningfully filter anyone out.

What ICP match rate should I be targeting?

There's no universal benchmark, but teams with a tightly scoped ICP and targeted demand gen commonly land in the 15 to 30% range of total identified traffic. Track the trend over time rather than chasing a specific number in isolation.

Identification rates measured against engaged sessions. Results may vary by traffic profile, geography, and industry.