B2B website visitor identification is the practice of matching an anonymous session on your site to a company — not a person — so sales can decide whether a visit deserves a follow-up. A useful implementation answers a narrow question with evidence: which organization’s network the request came from, which pages it read, how well that behavior matches the accounts you sell to, and how confident the match is.
Most teams buy the category expecting a name, an email and a reason to call. What they get is a company-level signal with gaps: home broadband looks like nobody, mobile traffic looks like nobody, and a shared office looks like six companies at once. That is the boundary of the method, not a defect in the product.
This guide separates what identification can prove from what it can only suggest, and shows how to turn a company visit into a sales signal your reps will act on. We build the capability into Prynt Reveal, so the examples reflect how the data behaves in production.
What visitor identification can and cannot resolve
The company, when the network belongs to the company
A match works by resolving the network behind a request to an organization: a business IP range, an ASN registered to a company, or an office network seen consistently. From that network, Reveal names the company with a confidence level and the evidence behind it — the resolved organization, the network type, the pages visited. Confidence is the part that matters. A single-tenant office range is stronger evidence than a range that also serves a coworking floor.
The person, only from your own first-party data
People are never reverse-engineered from anonymous traffic. A person links to a company only through first-party records you already hold — a form, a login, a CRM import — with a consent record attached. No data brokers, no guessed emails. That constraint is exactly what makes the output safe to put in front of a rep.
Where the signal breaks
Home broadband, mobile carriers, VPNs, Tor and datacenter traffic stay anonymous by design; the alternative is guessing, and guessed identities create bad CRM records and worse outreach. Two consequences shape how you use it.
- Most traffic is unidentifiable. Treat identified companies as the exception worth routing, not the whole plan.
- Shared networks blur. Agencies, incubators and coworking spaces resolve to the office, not the tenant. Read those sessions as “someone here is interested” and let the confidence field stop you from mass-emailing them.
Before you attribute a session to a company, know whether it is a bot, an AI agent or a masked proxy: that is a separate layer from identification, and a bot visit scored as intent will waste a rep’s afternoon.
Turning a company visit into a sales signal
Score intent from pages, not pageviews
Pageviews measure noise. Intent lives in a small set of pages read in a short window: pricing, security, integrations, docs, comparison pages, and repeat visits to the same high-value page. Build the score around questions a rep would ask — did they read pricing twice this week, did they open a page that only matters before a purchase, did the same account come back after a gap. Keep it explainable: a score you cannot trace to pages and a time window stops being trusted within a month.
Fit before enthusiasm
A visit from a company outside your ICP is not a lead, however deep the session. Score fit separately and escalate only when fit and intent are both present. Rule-based playbooks keep the volume manageable: one for “ICP account reading pricing,” another for “existing customer reading support docs,” which is a support signal, not a sales one.
Deliver where the rep already works
Routing beats dashboards. Push the company, intent, fit and evidence to Slack, HubSpot, Salesforce, n8n, Zapier or a signed webhook, and make delivery idempotent so a returning account updates one record instead of creating three. A drafted opener that a person reviews and sends is useful; anything sent automatically is not. A first playbook worth writing:
fit == high
and page in ("/pricing", "/security")
and visits >= 2 in 7 days
→ Slack #sales + HubSpot note, draft only
Privacy boundaries you have to respect
Company-level identification still touches personal data, and rules differ by jurisdiction. Make three things explicit: what you collect, why, and how someone opts out. Reveal ships a public opt-out, suppression lists, a privacy notice and a processing agreement to review before purchase — start with the privacy notice and the opt-out page, the two documents a legal reviewer asks for first. In the EU or UK, document legitimate interest, keep retention short, and never let a score become a reason to guess an identity.
FAQ
Does visitor identification give you email addresses?
Not from traffic. Identification resolves the organization behind a network; contact details come from your own first-party records. A tool that promises named contacts from anonymous visits is inferring, and inference produces bounced outreach.
How accurate is a company match?
It depends on the network and on whether the tool shows its evidence. A company’s own office network is strong; home broadband, mobile and VPN traffic is unidentifiable; shared offices are ambiguous. Judge a vendor by whether it shows confidence and reasons rather than a bare company name.
What if I only care about ten target accounts?
Filter hard. Keep a target list, score fit against it, and let the tool tell you when one of those companies returns. Ten routed visits beat a thousand unnamed pageviews, and the workflow costs you a rule, not headcount.
See it on your own traffic
The free plan of Reveal covers 100 companies a month on one site with 7-day history — enough to learn how much of your traffic is identifiable at all. Connect one site at pryntid.com and read a week of company visits before you decide anything else. If you are weighing options, our notes on how visitor identification approaches differ cover the tradeoffs.
Try it free
Prynt is device intelligence with a free plan: 5,000 identifications a month, visitor IDs with confidence and bot detection — no credit card. Upgrade to Pro for VPN, proxy, tampering and behavioral signals.