What Is Website Visitor Identification?

Most B2B website traffic arrives in analytics as sessions with no name attached. Visitor identification tools put a company name, and sometimes a person's name, on those sessions. Before buying one, it helps to know where that name comes from, how often it is right and what the law says about collecting it.

It works best on B2B sites where buyers browse from company networks and the team can act on account-level signals. Person-level matching adds names mainly for US traffic. A match is evidence of a visit, not a lead or consent to contact, and home, mobile and VPN traffic often resolves only to an internet provider.

What Is Website Visitor Identification?

Website visitor identification is the practice of matching an anonymous website visit to the company that owns the visitor's network and, in some tools, to a named person, so a B2B team can see which accounts are researching it before anyone fills in a form.

Take an illustrative example. A standard analytics tool tells you that someone from Manchester read your pricing page twice. A visitor identification tool tries to tell you that the visit came from a named logistics company, with its industry and size. A person-level tool goes one step further and tries to name the employee.

Those are two different outputs, and vendors often bundle a third:

• Company match: The organisation behind the visit, with firmographics such as industry, size and location.
• Person match: A named individual, usually with a job title, a LinkedIn profile and an email address.
• Intent signals: What the visitor did, such as pages viewed, return visits and time on high-value pages.

The difference matters because the methods differ. Leadfeeder (known as Dealfront until March 2026), which only does company-level matching, puts the limit plainly. A company-level tool matches the visitor's IP address to the company that owns it and has no way to know which employee was at the keyboard.

Most explanations describe identification in layers. A reverse IP-to-company lookup can name an organisation. Cookies, pixels or device identifiers, joined in an identity graph, can sometimes name a person, but that layer resolves a much smaller share of visits.

Whatever the layer, a match is evidence about a visit. It is not a lead, and it is not permission to contact anyone.

Anonymous Analytics Vs Visitor Identification

Anonymous analytics tells a team how many sessions came from which channels and pages, while visitor identification adds who was behind them, either a named company with its industry and size or, in person-level tools such as RB2B, a named individual with a LinkedIn profile and business email.

Standard web analytics is aggregated on purpose. It answers questions about channels, content and conversion paths, and it never needs to know who any one visitor is. Identification answers a different question, which is who to follow up with.

The company-match row reflects what tools like Leadfeeder return per company. Its guide lists firmographics (name, domain, industry, employee count, revenue band and location) alongside pages viewed, return visits and engagement.

The person-match row reflects what person-level vendors describe. RB2B says it sends names, LinkedIn profiles, business emails and company data to Slack, Teams and CRMs. CustomersAI says its B2B product returns personal email, mobile phone, address, LinkedIn profile, job title and seniority.

Moving down the table, the data gets more specific and the legal and accuracy questions get harder. That trade-off runs through the rest of this guide.

How Does Website Visitor Identification Work?

Website visitor identification works in layers, with a script on the site recording each visit, an IP-to-company lookup naming the organisation behind the network, and, in some tools, cookies, device IDs and an identity graph linking the browser to a named person before the record is sent to a CRM or Slack.

Here is the sequence a typical tool follows:

• Step 1. A script or pixel on the site records the visit, the pages viewed and the visitor's IP address.
• Step 2. The IP address is checked against IP-to-company data built from public registrations and network data.
• Step 3. Cookies, device IDs and third-party IDs in an identity graph refine the match or extend it to a person.
• Step 4. The match is enriched with firmographics or contact details from a company or contact database.
• Step 5. The record is delivered to a CRM, Slack, Teams or ad audiences for someone to act on.

Step 1, the script: Installation is light. Leadfeeder describes a JavaScript snippet deployed directly or through Google Tag Manager or Segment, usually in under 30 minutes. RB2B installs as one script in the site header.

Step 2, the IP lookup: Leadinfo says it links an IP address to a company through public registrations and network data, then compares it with a company database. This is the layer that works without naming anyone.

Step 3, the identity graph: Demandbase describes an identity graph of cookies, IP addresses, third-party IDs, device IDs and public VPN and ZTNA providers, plus machine-learning predictions. Leadfeeder's guide adds that first-party cookies let tools link sessions and recognise a returning visitor when the IP address changes.

Step 4, enrichment: Leadinfo says a separate contact database of more than 300 million profiles supplies decision-maker emails, phones and LinkedIn profiles after the company match. That contact data comes from a database lookup, not from the visit itself.

Step 5, delivery: Leadfeeder lists CRM sync with Salesforce, Pipedrive, Microsoft Dynamics, HubSpot and Zoho, Slack alerts and ad-audience sync to LinkedIn, Google and Meta.

The weak point is step 2. An office network has a registered owner, but a home broadband line, a mobile network or a VPN usually does not belong to the visitor's employer. That traffic often resolves only to an internet provider. Demandbase says its probabilistic models may deviate because of dynamic IP usage and remote work, and that iOS and cellular traffic is harder to identify.

What Is Company Matching?

Company matching is the step where a visitor's IP address is looked up against databases of registered IP blocks, so a visit from a corporate network resolves to that company's name, industry and size, while a visit from home broadband or a mobile network often resolves only to the internet provider.

Companies and network operators register blocks of IP addresses. Vendors collect those registrations, add their own network data and map each block to a company record. The size of that map is a selling point, and two vendors describe theirs in different terms:

• Leadfeeder says it holds more than 45 million IP-to-company mappings covering about 9 million companies.
• Demandbase says it maps around 3.7 billion IP addresses to a company or otherwise classifies them, supplemented by hundreds of millions of active cookies.

These are each vendor's own scale claims on its own basis. They are not a like-for-like measure of coverage, and a bigger number does not mean more of your visitors will match.

Company matching has three limits worth knowing before a trial:

• Remote work: An employee on home broadband shows up as the internet provider, not the employer.
• Shared IPs: Demandbase says shared IP addresses, for example in India, limit regional coverage.
• Mobile traffic: iOS and cellular traffic is harder to tie to a company.

Forms correct the guess. Demandbase says that when a visitor types a business email domain into a form, the company from that domain replaces the IP-identified company. A consumer email domain leaves the IP match in place. Known visitors should always outrank IP guesses in your CRM.

The skeptic's point is simple. A company match names the owner of a network, which is not always the buyer's employer. Treat it as a strong hint about an account, not a fact about a person.

If company-level data is what your team needs, the guide to company-level visitor identification tools compares the options on account recognition and match evidence.

For larger teams with account hierarchies and governance needs, there is a separate comparison of enterprise visitor identification tools.

Company-Level Vs Person-Level Visitor Identification

Company-level identification tells a team which organisation visited, using IP-to-company data that works worldwide, while person-level identification names the individual through cookies, device IDs and publisher networks, and vendors such as RB2B limit it to US traffic because of EU and UK privacy law.

This is the real decision behind choosing a tool. The two levels use different signals, cover different markets and carry different legal weight.

Company level: Leadfeeder and Leadinfo work here. Leadinfo says it shows only company name, sector, size and location. Leadfeeder says it does not do person-level identification at all.

Person level: RB2B says it matches visitors against a fully permissioned publisher network using first- and third-party cookies, device IDs and IP addresses. It says person-level ID resolves only US traffic, and its global company-level ID runs through a partnership with Demandbase. CustomersAI goes further on the data it returns, including personal email and mobile phone.

A practical decision rule follows from those facts:

• Start company level if you sell into the EU or UK, or if your motion is account-based and reps work accounts rather than individuals.
• Consider person level if you sell to US buyers and reps work individual leads. It adds names, but it needs a sharper review of legal basis and data quality.
• Combine them only when you can keep the two data sets apart by region in your CRM.

Person-level tools also differ in what happens after the match. For one worked example of that difference, see the comparison of RB2B vs Warmly.

How Accurate Is Website Visitor Identification?

Website visitor identification accuracy depends on what is being counted, since vendors quote match rates against total traffic, B2B traffic or identified visitors, and bots make up over 53% of all web traffic according to Imperva's 2026 Bad Bot Report, which skews any rate measured on raw sessions.

Every published match rate has an owner and a denominator. Here they are side by side, each kept on its own basis:

Read the table row by row, not as a ranking. A few points stand out.

Leadfeeder's two figures are different things. The 10–40% and 5–20% ranges are Leadfeeder's summary of independent testing. The 38.0% figure is Leadfeeder's own result across its five core regions in August 2026.

Demandbase gives two denominators. It says a client-side cookie integration identifies about 20% of B2B traffic, which is roughly 10% of all site traffic. The same rate looks halved when the denominator widens.

RB2B's figure has no stated method. It claims 30–40% of visitors identified at person level in the US, and 70–80% when person- and company-level are combined.

CustomersAI's figure comes from an unnamed tester. It reports 65–85% data accuracy from a blind test by an unnamed third-party CDP, based on tests with US DTC brands from February to June 2025. It also claims most visitor ID solutions get identities right 5–30% of the time, with no named source.

None of these figures can be averaged or compared as like for like, and none of the vendor figures is independent.

The most reliable test is your own data. CustomersAI describes testing accuracy by comparing predicted identities against form-fill submissions, and that method works for any tool. Take the visitors who filled in a form with a business email, then check whether the tool had already named the same company or person.

Is Website Visitor Identification Legal Under GDPR And CCPA?

Website visitor identification is legal in the EU and California only on the right basis, because the Court of Justice of the EU held that a visitor's dynamic IP address is personal data when the site operator can lawfully identify the person, and California's regulator treats IP addresses as personal information too.

EU: The ruling came in Case C-582/14, Breyer, in October 2016. Because a dynamic IP address can be personal data for the site operator, even company matching processes personal data in many setups. Company-level does not mean outside the GDPR.

California: California's privacy regulator lists device identifiers such as an IP address or mobile advertising ID as personal information under the CCPA and the Delete Act. Deidentified data is not personal information.

Vendors then take their own legal positions. These are positions, not rulings:

• Leadinfo argues company data is not personal data, relies on GDPR Article 6(1)(f) legitimate interest, and says its cookieless method needs no ePrivacy cookie consent. It offers company opt-out within 48 hours.
• RB2B ringfences person-level ID to US traffic, says its database is designed to exclude EU and UK residents' personal identifiers, and says customers must use a consent management solution if they deploy its global company-level ID. Its own page recommends consulting a privacy attorney.
• Leadfeeder says person-level identification in the EU has to clear two separate legal hurdles.

"Cookieless" is often presented as the whole answer. It answers the cookie-consent rule, but it does not settle whether processing an IP address needs a lawful basis. Under the Breyer reasoning, it often does.

In practice, settle three things before installing a script. Decide the lawful basis for each region, decide whether your consent tool must load before the identification script, and record which data you keep and for how long.

This section is general information, not legal advice.

Which Tools Identify Website Visitors?

The five tools split by what they name, with Leadfeeder, Leadinfo and Demandbase matching visits to companies, and RB2B and CustomersAI adding named people, which makes the choice of identification level the first decision rather than the vendor. These are examples of each approach, not a ranking.

Leadfeeder fits teams that want company-level account lists in the CRM. It returns firmographics, pages viewed and return visits, with ICP filters and alerts on pages such as pricing. It does not name people.

Leadinfo fits the same company-level job and keeps its data on EU servers in Ireland and Frankfurt. It says it recognises more than 220 million companies from over 195 countries, and lists integrations with Salesforce, HubSpot, Pipedrive, Microsoft Dynamics, Zendesk Sell, Teamleader, Microsoft Teams and LinkedIn Ads. Its price tiers are based on unique companies identified per month.

RB2B fits US-focused teams whose reps work individual leads from Slack alerts. It sends names, LinkedIn profiles and business emails to Slack, Teams, CRMs and Clay. Its person-level ID covers US traffic only, and it bills per resolution, so cost tracks traffic and match rate.

CustomersAI fits ecommerce-led teams that want person-level data feeding email and ad audiences. It identifies visitors through its X-Ray Pixel and sends data to Klaviyo, Salesforce, HighLevel and Sendlane, and to Meta and Google audiences. Its accuracy claims rest on an unnamed tester and US DTC brands.

Demandbase fits enterprise teams that want account identification inside a wider ABM platform. It identifies accounts before any form fill and replaces IP guesses with the company from a business email domain. It does not show IP addresses for identified accounts.

Once you know which level you need, the shortlist of the best website visitor identification tools compares the options head to head.

How To Implement Website Visitor Identification

Implementing website visitor identification starts with choosing company-level or person-level matching for the markets you sell into, then settling consent, installing one script, excluding internal traffic and routing each match to a named owner in the CRM or Slack.

Illustrative scenario: A 40-person SaaS company sells into the UK and the US. It has one marketer, three account executives and HubSpot as its CRM. These are assumptions for the example, not a case study. Its UK traffic needs a company-level, lawful-basis-first setup, and its US traffic could support person-level alerts.

• Decide the identification level by market. Company level for UK and EU traffic, person level only for US traffic if reps will act on names.
• Settle the lawful basis and consent setup before installing. Document legitimate interest or consent per region, and check whether the vendor requires a consent tool.
• Install the script through your tag manager. Leadfeeder supports Google Tag Manager and Segment, and RB2B uses one header script.
• Exclude your own staff, partners and low-value pages. RB2B's Page Suppression excludes traffic by domain, URL or list.
• Set ICP filters and alerts on high-intent pages such as pricing, so reps see fit accounts first.
• Route every match to the CRM or Slack with a named owner and a follow-up rule.
• Plan for the form override, so a business email from a form replaces the IP guess on the record.

In the scenario, the company would turn on company-level matching for all traffic, route UK matches to the CRM as account records and send only US person-level alerts to Slack. The marketer would own exclusions and filters, and each account executive would own the accounts in their territory.

The biggest planning mistake is cost. Billing scales with what the tool finds, not with seats:

• RB2B bills per resolution, one identified visitor delivered. Plans include a monthly number of resolutions with per-resolution overage of 45 cents on Starter and 25 cents on Pro and Pro+. Extra domains are a paid add-on on Pro and Pro+.
• Leadinfo bases its tiers on unique companies identified per month, which it estimates at roughly 30% of website visitors.

So traffic growth, a campaign spike or a bot surge can raise the bill. Excluding internal and low-value traffic is a cost control as much as a data-quality step.

How To Measure Visitor Identification Success

Visitor identification is working when a growing share of matched companies fit the ideal customer profile and reach a sales conversation, and when matches checked against the team's own form fills hold up, not when the raw count of identified visitors rises.

Raw identification counts flatter every tool. The measures below show whether the data is useful:

ICP match share uses the tool's own filters. Leadfeeder describes ICP filters on firmographics, so the share of fit companies is easy to pull from the dashboard.

Verified match rate is the form-fill test. Compare the company or person the tool named against the business email domain the visitor later typed in. Demandbase's form override shows the logic, since a business domain is stronger evidence than an IP match.

Clean traffic share matters because over half of web traffic is bots, per Imperva. Remove bots and your own staff before calculating any rate, using exclusions such as RB2B's Page Suppression.

Cost per useful match follows the billing unit. For RB2B that is resolutions, and for Leadinfo it is identified companies per month. Divide spend by the matches that fit your ICP, not by all matches.

Sales conversations are the end measure. Count identified accounts that reach a meeting within the trial window, and compare them with accounts that arrived through forms alone.

The decision rule is to judge a trial on the ICP matches a rep acts on, not on raw identifications. A tool that names fewer companies but more of the right ones is the better tool for your team.

Frequently Asked Questions

Sources