Website Visitor Identification Accuracy: What to Expect

Website visitor identification accuracy has two parts: how many visits get a name, and how often that name is right. Company-level tools are commonly quoted at roughly 20 to 35 percent of B2B traffic named, with accuracy that varies by audience, so verify a sample yourself.

Coverage and correctness are different numbers

Vendors usually advertise coverage, also called match rate or fill rate. That is the share of visits that received any company label. Correctness is the share of those labels that were true and useful. A tool can have impressive coverage and poor correctness if it labels internet providers as customers.

The figures quoted in this article come from BusinessMCP’s pages, which are vendor-published. They state company-level resolution of roughly 20 to 35 percent of B2B traffic and person-level resolution of roughly 5 to 20 percent, US only, where enabled. BusinessMCP also states that about 98 percent of visitors never fill out a form and about 2 percent do. Treat these as the vendor’s view, though they are consistent with how the technique behaves.

What moves the number up or down

Factor	Effect on match rate	Why
Audience size of the employer	Higher for mid-market and enterprise	They run their own network ranges
Remote work	Lower	Home broadband names the provider
Region	Higher in North America and Western Europe	Better registry data
Mobile share	Lower	Carriers resolve to the carrier
Consumer or SMB focus	Much lower	Few own networks
Cloud and bot traffic	Inflates raw matches	Resolves to hosting firms
VPN use	Lower	Resolves to the VPN provider

BusinessMCP’s website visitor identification page acknowledges the same pattern: remote workers, mobile and VPN users often resolve to providers rather than employers.

Key takeaway: the right question is not “what is your match rate?” but “what is your match rate on traffic that looks like mine, after removing providers and cloud hosts?”

Why price rarely buys accuracy

BusinessMCP’s small-tools roundup claims that identification accuracy depends on the underlying IP-to-company data, and that budget tools license much of the same data as expensive ones. If that is right, you should expect roughly similar coverage from different price tiers, with differences in features around it. We cannot verify the licensing claim independently, but it supports a practical stance: test, do not assume that paying more names more.

A repeatable test you can run in a week

1. Choose the pages. Track the same set in every tool: pricing, product, comparison and contact.
2. Run tools in parallel. Install two trackers on the same pages. Do not change traffic sources during the test.
3. Export the named results. Take company name, domain and visit count.
4. Remove non-targets. Delete internet providers, mobile carriers, cloud hosts, your own staff and known crawlers.
5. Sample ten. Pick ten at random from each tool and check whether the company plausibly belongs to your market, using their website and one other signal.
6. Score. Count correct and relevant results divided by total named. That is your precision. Then count relevant accounts found by each tool: that is your useful coverage.

The same procedure works whether you are comparing a new vendor or deciding whether to renew. Our switching checklist shows how to run it during a migration.

Red flags in accuracy claims

• A single match rate with no traffic mix.
• Coverage figures that include internet providers.
• No distinction between company and non-business verdicts.
• No stated data refresh cadence.
• Global person-level identification, which the BusinessMCP guide flags as a red flag, and which is a legal risk in the EU.
• Comparison tables with no limits listed for the vendor itself.

BusinessMCP’s guide says near-100 percent claims involve overpromising, padding with stale data or counting ISP hits as companies. We agree with the substance, whichever vendor is making the claim.

Improving accuracy on your side

You cannot change how a registry maps addresses, but you can raise the quality of what you act on.

• Capture more email domains. Forms, gated content and tracked email links give deterministic matches. See how to follow up on website visitors.
• Combine signals. A company that visits pricing twice in a week is more meaningful than one that glanced at a blog post. Lead scoring from website visits explains how.
• Filter early. Set exclusion rules for providers, bots and employees so noise does not reach your sales team.
• Review monthly. Spot-check a sample each month to catch drift.

Setting expectations with your team

Tell sales plainly that roughly one visit in three to five at best can be tied to a company in a typical B2B audience, and that a name is a hypothesis. Teams that expect a contact list get disappointed. Teams that expect a prompt to investigate get value.

A realistic target is a small set of high-fit accounts per week that you would not otherwise have noticed. If a tool delivers that, its accuracy is good enough for the job, whatever the percentage on its homepage says.

How to put the number in a business case

Never present a single match rate to finance. Present a range with the assumptions behind it. For example: monthly business-page visitors, a low and high match rate from the commonly quoted 20 to 35 percent band, a filter for providers and repeat visits, and an estimate of the share of named accounts that fit your profile. Multiply through for a low, central and high case, and show what the tool costs per relevant account in each.

If the low case still justifies the spend, you have a robust argument. If only the high case does, run a trial first. For a flat-priced option that can be tested on a free plan, see BusinessMCP’s Leadfeeder alternative comparison, and remember that it too quotes the 20 to 35 percent range for any vendor.