“What match rate should I expect?” is the first question every buyer asks me. The honest answer for years was a shrug. Vendors quote whatever marketing number sounds impressive that quarter. Buyers quote whatever their last tool showed before they churned. Most of the published numbers are either company-level identifications dressed up as person-level, or one cherry-picked customer dressed up as an average.
I am George, founder of Leadpipe. We run an identity graph behind 280M verified profiles, 5M websites monitored, and 60B intent signals refreshed every 24 hours. That gives us a real vantage point on what match rates look like in production, against live US B2B traffic, on a deterministic person-level basis.
This post is the honest version of “what to expect.” It is not “we ran a study and the median was 31.4%.” There is no universal median, because match rate is a function of your traffic composition, not just your vendor. What I can give you is the verified Leadpipe baseline, the structural reasons match rates vary, and a methodology to measure your own.
The verified baseline
There is one number I will commit to, on the record, and it is the only one I can verify across our network:
Leadpipe identifies 30-40%+ of US B2B website visitors at the person level, deterministically, with full contact data.
Above 40% is reachable on favorable traffic mixes. Below 30% happens on traffic mixes the identity graph cannot resolve well. The 30-40%+ band is the typical operating range for US B2B traffic running our pixel.
In an independent accuracy test on 75,000 visitors over 120 days, Leadpipe scored 8.7/10 on identification accuracy, against RB2B’s 5.2/10 and Warmly’s 4.0/10. That test measured “of the people you identified, how many were correctly identified.” It is a different question from match rate. Match rate is “of total visitors, how many did you identify at all.” Both numbers matter. They measure different things.
Independent accuracy test (75,000 visitors, 120 days):
Leadpipe ████████████████████ 8.7/10
RB2B ███████████ 5.2/10
Warmly ████████ 4.0/10If you are evaluating vendors, ask both questions: what is your match rate on US B2B traffic, and what is your accuracy on the people you identify. Vendors who can only answer one of those are filling in the other from marketing copy.
Why “the average match rate” is a bad question
The thing buyers want is a single number. The thing the data shows is a distribution that is dominated by traffic composition.
A match rate is computed as identified unique visitors divided by total unique visitors, per site, per month. The denominator (total visitors) is what most teams under-think about. Two sites with identical traffic volume can have wildly different match rates because the composition is different.
| Composition factor | Effect on match rate |
|---|---|
| Geography (US vs international) | US identity graph coverage is deepest; international traffic resolves at lower rates |
| Audience type (B2B decision-makers vs general consumers) | B2B-shaped audiences match higher; consumer-shaped traffic dilutes the rate |
| Traffic source mix | LinkedIn organic over-indexes; some paid social and broad-reach campaigns under-index |
| Bot share in your analytics | Bot sessions in the denominator artificially depress your match rate |
| Return-visitor share | Match rate climbs with repeat visits as more signals accumulate |
| Device mix | Mobile traffic resolves at lower rates than desktop, though the gap is narrowing |
Every one of those is a denominator effect. Same vendor, same identity graph, two different sites, two very different match rates.
The implication is that “what match rate should I expect” without context is unanswerable. With context, it is answerable in a tight band.
What drives the bottom quartile
Sites that consistently land in the bottom of the match-rate distribution share three traits.
Heavy international traffic. Identity graph depth is highest in the US, moderate in Western Europe, thin elsewhere. If 60% of your traffic is outside the US and EU, your match rate on the total-visitor denominator will be lower even if your US-slice match rate is excellent. Look at the rate by country before you blame the tool.
Consumer-shaped audience. B2B sites that attract a lot of curious individuals (free-tool SEO plays, viral content, broad-reach paid social) get clicks from people who are not decision-makers at identifiable businesses. The traffic counts in the denominator. The matches do not.
Bot-contaminated analytics. If your analytics tool is counting bot sessions in the denominator, your match rate looks artificially depressed. The full topic is in our bot traffic study. Filter bots before you compute the rate.
If any of those describe your site, the match rate is doing what the math says it should. The fix is not switching vendors. The fix is either filtering the denominator or accepting the rate.
What drives the top quartile
The mirror image. Sites that consistently land in the top of the distribution share three traits.
US-focused traffic. Often >80% US visitors. The deepest identity graph coverage is in the US, and concentrated traffic there compounds match rates.
Intent-driven content strategy. Pages that pull in-market buyers (comparisons, pricing, integrations, “alternatives to” posts) attract decision-makers, not generalists. The audience composition tilts toward identifiable B2B people.
Return-visitor share. Match rate climbs with repeat visits as the identity graph accumulates more signals across sessions. Sites that drive return traffic through email, retargeting, and product-led content match better.
A 40%+ site-level match rate is reachable with those levers in place. It is not the average. It is the upper end of the distribution. Do not expect it if your mix is global and top-of-funnel.
How to measure your own honestly
Concrete framework. None of this is vendor-specific.
Step 1: Filter your denominator
| Filter | Why |
|---|---|
| Remove bot traffic | Otherwise bots dilute the denominator |
| Remove internal IPs | Your team is not a buyer |
| Remove sub-30-second bounces from non-organic | Often non-human or low-intent click-throughs |
| Slice by country | US-only is the right comparison band for US-trained graphs |
If you skip this step you are computing a number that has no diagnostic value. A site with a “13% match rate” might have a 32% rate on US human traffic and a 60-bot-share on an international long-tail.
Step 2: Run for at least 30 days
Match rates climb as the identity graph accumulates signals. The first week of a fresh pixel install systematically under-reports compared to month two. Run a clean 30-day window before drawing conclusions.
Step 3: Compute site-level, not pooled
If you are reading vendor benchmarks, prefer site-level distributions over pooled visitor-level averages. Pooled averages are dominated by the largest sites in the sample. The site-level distribution is a more useful signal for your own expectations.
Step 4: Compare against your industry
A 28% match rate on healthcare IT is doing well. A 28% match rate on US SaaS is below average. The industry benchmark report shows the typical bands across 12 sectors. Read your number against your industry, not against the headline.
Step 5: Distinguish person-level from company-level
This is the biggest source of confusion in the category. A vendor that quotes “70% match rate” on company-level identification (knowing which company visited) and a vendor that quotes “30-40%+” on person-level identification (knowing the actual individual) are not measuring the same thing. Person-level is the harder problem and the more valuable answer. The person-level vs company-level explainer walks through the difference.
| Identification level | Typical match rate | What you get |
|---|---|---|
| Company-level only (IP reverse lookup) | 50-70% on mixed traffic | ”Someone from Acme Corp visited” |
| Person-level (deterministic graph) | 30-40%+ on US B2B (Leadpipe baseline) | “Sarah Chen, VP Marketing at Acme Corp, visited your pricing page” |
| Person-level (probabilistic / LinkedIn-only) | Variable (5-20% on RB2B traffic profiles) | Lower-confidence person guesses |
Match rate is not the only thing that matters. The accuracy on the people you identify matters as much. Probabilistic identification can produce a higher headline match rate by guessing aggressively, with a much lower correct-ID rate. The independent test results are the way to triangulate that, not the match-rate line on the pricing page.
Comparing vendor claims honestly
The visitor identification category has a truth-in-advertising problem. Match-rate claims from various vendors range from “40% of US B2B visitors” to “up to 70%” to “90%+” in a few cases. Most of those numbers are either talking about company-level identification, which is an easier problem, or quoting their single best customer.
Three questions to ask any vendor:
- Is that match rate person-level or company-level?
- Is it deterministic or probabilistic?
- Is it the median across customers, or a single customer’s result?
If a rep cannot answer those three, the number they gave you is not useful. Compare it to the verified Leadpipe baseline of 30-40%+ on US B2B, person-level, deterministic, averaged across customers, and ask why their number is so much higher.
For where this fits into a broader buyer-evaluation framework, see top 10 visitor identification softwares and the accuracy independent test.
Implications for the reader
Three things to do with this.
Recalibrate your expectations. If you are evaluating vendors and a sales rep quotes 70% match rate, ask whether that is person-level, deterministic, US-only, and averaged across customers or quoted from one. Compare their answer to the verified Leadpipe baseline.
Diagnose where you sit. Pull your own match rate, after filtering the denominator. Compare it to the industry benchmark for your sector. If you are below, check traffic mix first (geography, source, bot share) before you blame the tool.
Size the gap, not the rate. A 28% person-level match rate on 100K monthly visitors is 28,000 named people per month. That is the number that matters for pipeline. The visitor-to-conversion gap study shows what those identified visitors are worth in pipeline dollars. Match rate is the leading indicator. Pipeline created from identified visitors is the lagging indicator. Track both.
What 2026 changes
A few structural shifts worth flagging if you are setting expectations for the year:
| Trend | Effect on match rate |
|---|---|
| AI agent traffic (Perplexity, ChatGPT, agent browsers) | Adds noise to the denominator, often unidentifiable |
| Tightening browser privacy defaults | Probabilistic methods get worse; deterministic graphs unaffected |
| LinkedIn-only identification limits | Methods that depend on LinkedIn login state see drift |
| Identity graph freshness (24-hour vs weekly refresh) | Daily-refresh graphs hold their match rate; stale ones decay |
Leadpipe refreshes the graph every 24 hours and uses deterministic matching against first-party signals, which is why our baseline holds 30-40%+ on US B2B traffic in 2026. Methods that depend on probabilistic guessing or weekly batch refresh have lost ground over the last year.
Limitations of any benchmark
Worth stating clearly:
- One vendor’s graph. Leadpipe’s match rate is what Leadpipe’s identity graph produces. Another vendor with a different data footprint will produce different numbers. The distribution shape is similar; the absolute numbers are not portable.
- US B2B focus. The 30-40%+ band is the US B2B baseline. International benchmarks are lower. EU/UK runs company-level by default with person-level requiring affirmative consent.
- Person-level, deterministic. The numbers in this post are person-level, deterministic, full contact data. Company-level-only and probabilistic comparisons are different categories.
- Match rate is not accuracy. A high match rate with low accuracy is worse than a moderate match rate with high accuracy. Read the independent accuracy test alongside any match-rate number.
The point of the benchmark is not to produce a single national average to compare yourself against. It is to set honest expectations, give you a methodology for measuring your own, and stop the worst of the vendor-marketing inflation in the category.
Leadpipe identifies 30-40%+ of your US B2B visitors with full contact data on the Pro plan at $147/mo. No credit card to start the 500-lead trial. Start identifying visitors →
