Most B2B teams already have website visitor tracking. Google Analytics is installed, a tag manager fires a dozen events, and a dashboard shows sessions going up and to the right. Then a sales rep asks a simple question, "Which of my accounts were on the site this week?", and nobody can answer it. That gap is the whole problem. Visitor tracking built for traffic reporting tells you how many people came. B2B visitor tracking has to tell you who came, what they cared about, and whether it is time to act.
This post covers what B2B website visitor tracking should capture, the three layers of tooling that do it, an event spec you can copy into your tag manager, and the privacy rules that decide what you are allowed to collect in the first place.
Why Pageview Tracking Fails B2B Teams
Consumer analytics assumes lots of visitors, short journeys, and a conversion in the same session. B2B looks nothing like that. A deal involves a committee, the research spreads over weeks, and most of it happens before anyone fills in a form.
Our 2026 B2B personalization research found that 73% of buying committees visit a vendor's website at least 3 times before the first sales conversation, and on enterprise-focused sites, the fourth visit is the most common moment for a first form fill. Session-based analytics records each of those visits as an unrelated session from an anonymous user. The pattern that matters (one account, three people, five visits, pricing page twice) is invisible.
Three specific things break:
- The unit of analysis is wrong. GA4 counts users and sessions. Your revenue team works in accounts. Without an account layer, you cannot join website behavior to pipeline.
- Anonymous traffic is a black hole. Only 5-10% of visitors ever fill out a form, so 90% of your traffic never gets a name attached. If tracking only gets useful after a form fill, it ignores most of your demand.
- Events are tracked for reporting, not for action. "Scroll depth 75%" is fine for a content report. It does not tell a rep or a personalization rule anything.
The Three Layers of B2B Website Visitor Tracking
Good B2B tracking stacks three layers. Each answers a different question, and you need all three before the data is worth routing anywhere.
Layer 1: Traffic analytics (how many, from where)
This is GA4, Plausible, Matomo, or whatever you already run. It answers channel and content questions: which campaigns drive visits, which pages hold attention, where people drop off. Keep it. Just stop expecting it to answer account questions. GA4 deliberately avoids identifying organizations. Google states that Analytics does not log or store individual IP addresses from EU, Swiss, or UK users, using them only to derive coarse location before discarding them.
Layer 2: Company identification (who)
This layer resolves anonymous sessions to companies, mostly through reverse IP lookup against business IP ranges, then enriches the company with industry, size, and location. Coverage is partial by design. The median B2B site on our platform identifies 32% of traffic at the company level, with enterprise-heavy sites reaching up to 65% and SMB-heavy sites closer to 15-25%. Remote work and consumer VPNs are the main reasons it is not higher. Our explainer on how company-level visitor identification works covers the mechanics and accuracy limits in detail.
Layer 3: Behavioral events (what they cared about)
The third layer records the specific actions that signal intent, tied to the identified company. Not every click. A short list of events that a person on your revenue team would actually react to. This is where most teams over-collect. Forty events nobody looks at are worse than eight that drive a rule or an alert.
What to Track: A Copyable Event Spec
Here is the event spec we recommend as a starting point. It is deliberately short. Every event on it either feeds a score, triggers a personalization rule, or lands in a CRM timeline.
| Event | Fires when | Key properties | Used for |
|---|---|---|---|
pricing_viewed | Pricing page loads | plan tab opened, time on page | Intent score, sales alert |
case_study_viewed | Any customer story loads | story industry, story company size | Industry inference, proof personalization |
integration_viewed | Integration or docs page loads | integration name | Technical evaluator signal, tech stack inference |
comparison_viewed | "vs" or alternatives page loads | competitor named | Late-stage signal, competitive messaging |
demo_cta_clicked | Any demo CTA clicked | CTA location, CTA variant | CTA testing, conversion path analysis |
form_started | First field focused | form ID | Form abandonment recovery |
form_submitted | Successful submit | form ID, email domain | Identity stitching, CRM sync |
return_visit | New session from a known company within 30 days | days since last visit, session count | Engagement recency, score decay |
Notice what is missing: scroll depth, outbound clicks, video quartiles, time on blog posts. Track those in your traffic analytics if content reporting needs them. Do not send them into account scoring. They add noise and make every account look engaged.
Two properties matter on every event: a company identifier (from layer 2) and a persistent first-party visitor ID. With both, you can later merge an anonymous history into a known contact when that person submits a form, which is the moment the first three visits suddenly become useful to sales.
Turning Tracked Visits into Account Signals
Raw events are not signals. A signal is an event pattern that a human or a rule should act on. The step most teams skip is writing those patterns down. Start with four:
- New ICP account, first visit. An identified company that matches your ICP and has never visited before. Action: start showing industry-specific proof on their next pageview.
- Multi-person account. Two or more distinct visitors from the same company in seven days. Action: flag to the account owner. A committee forming is a stronger signal than one person reading a lot.
- Late-stage page cluster. Pricing plus a comparison or integration page in the same week. Action: raise the account's priority and swap homepage CTAs from content offers to a meeting.
- Re-engagement. A closed-lost or dormant account returns after 60+ days of silence. Action: notify the previous owner with the pages viewed.
Each of these maps to a segment you can personalize against. Our guide to behavioral segmentation for B2B websites goes further into how to turn these patterns into scoring rules without drowning in edge cases.
A Tool Stack That Covers All Three Layers
You do not need a dozen tools. A typical stack for a mid-market B2B team looks like this:
- Tag manager: Google Tag Manager or a server-side container, so the event spec lives in one place.
- Traffic analytics: GA4 or a privacy-first alternative like Plausible or Matomo for channel and content reporting.
- Company identification and personalization: one platform that resolves companies and can act on them on the page. Splitting identification and personalization across two vendors usually means two different company matches for the same visit.
- CRM: HubSpot or Salesforce, receiving account-level activity, not raw event streams.
- Alerting: Slack or email notifications for the four signals above, routed to the account owner.
Markettailor's visitor identification covers layers 2 and 3 in one script: it resolves the company, records the events above against it, and makes the same account data available to personalization rules on the next pageview.
Privacy: What You Can and Cannot Track
Company-level tracking and individual tracking are different legal questions, and mixing them up is how teams get into trouble.
- Cookies and local storage need consent in the EU and UK unless strictly necessary. The UK ICO's guidance on cookies and similar technologies says you must get consent to store a cookie on a user's device, and that cookies essential only for your own purposes still require it. A persistent visitor ID for analytics falls on the consent side.
- Company-level identification from a business IP range describes an organization, not a person, which keeps it out of most personal-data obligations. The moment you join it to a named contact, GDPR applies in full.
- Disclose it. A plain sentence in your privacy notice explaining that you personalize by company does not hurt conversion. Our research found that customers who published a personalization disclosure saw no measurable drop in conversion rates.
For the full breakdown, read our post on B2B personalization and data privacy compliance.
The Mistake We See Most: Tracking Without an Owner
The failure mode is rarely technical. Teams install identification, collect three months of account data, and then nobody owns the output. Marketing assumes sales is checking the dashboard. Sales assumes marketing will send alerts. The data ages out.
Before you add a single event, decide who receives each of the four signals and what they do within 24 hours. If you cannot name the person, do not track the event yet. Tracking is cheap. Attention is not.
Where to Start This Week
- Audit your current tag manager and delete events nobody has used in a report in 90 days.
- Add the eight events above with a company identifier on each.
- Write down the four account signals and assign an owner to each.
- Pick one signal (late-stage page cluster is the easiest win) and wire it to a Slack alert.
- After 30 days, check how many alerted accounts turned into meetings, and tune the thresholds.
If you want to see which companies are on your site today and what they are reading, our pricing page has a 20-minute walkthrough where we run identification on your real traffic.