Home › Guides › Reading intent signals

How to Read Intent Signals from Your Own Website Traffic

Published September 5, 2026 · BuyerJourney.net

Your traffic is already full of intent signals — pricing visits, comparison sessions, accelerating returns. The problem isn't detecting them; it's that in most setups they're fragmented across anonymous sessions, so nobody accumulates a readable score and nothing is actionable. Here's what the signals are, a scoring model simple enough to actually run, and the two fixes that stop the signal from going to waste.

The signal hierarchy

Two readings matter more than any single row:

• Compounding beats magnitude. One pricing visit is ambiguous — plenty of first-time visitors check cost out of curiosity. Three medium signals from the same person inside a week is a stronger buy signal than any single high-intent page.
• Context reorders the table. A pricing visit as a first touch is a low-grade signal; the same visit after two comparison sessions is decision-stage behavior. Sequence is the multiplier — the same principle as funnel-stage reading.

A scoring model you'll actually maintain

Skip the machine-learning ambitions and run something legible:

1. Points per behavior: pricing visit 3 · demo/trial page 3 · comparison page 2 · repeat visit within 7 days 2 · integration/security page 2 · educational page 1.
2. Sum per visitor over a rolling 14-day window — recency is the point; intent decays fast.
3. One threshold (say, 6+) that means "in-market now," reviewed monthly against what actually converted.

The uncomfortable truth about scoring: the model's sophistication matters far less than what it's applied to. A crude score over connected, person-level behavior outperforms an elaborate score over fragmented anonymous sessions — because in the fragmented case, the same buyer's signals are split across three "visitors" who each stay under threshold.

Why the signal goes to waste — and the two fixes

Most analytics setups leak intent in two places:

• The signals aren't connected. Cookie-scoped analytics splits one buyer across devices and days, so scores never accumulate. Fix: identity — sessions resolving to one verified person. BuyerJourney.net does this deterministically for 75–85% of matched traffic, and retroactively stitches earlier anonymous sessions once someone is identified, so the score inherits the history.
• The signals aren't named. Even a perfect score attached to an anonymous cookie gives sales nothing to do. A score attached to a verified individual — with the journey behind it — is a call list ordered by readiness. (Where this sits among tool categories: buyer journey tracking tools.)

The intent you can't see: off-site signals

On-site intent is the late intent — by the time someone reads your pricing page, they've usually researched the category elsewhere first. That off-domain research is invisible to your analytics but measurable: BuyerJourney.net connects 35B+ daily intent signals from across the open web into its journey data, so in-market behavior registers before the first visit and the journey opens at the true beginning. Turning that off-site layer into targetable ad audiences is the job of Intent Audiences, the suite's targeting side.

FAQ

Which behaviors count as intent signals?

High: pricing, demo, contact, post-comparison case studies. Medium: comparison pages, fast repeat visits, deep sessions. Low: single blog visits and bounces. Compounding medium signals beat any single high one.

How do I score intent?

Points per behavior, rolling 14-day sum per visitor, one in-market threshold reviewed monthly. Legibility beats sophistication.

Why does intent go to waste?

Fragmentation (one buyer split across anonymous sessions, so scores never accumulate) and anonymity (a hot score with no name attached). Identity resolution fixes both.

What about pre-visit intent?

Off-site research precedes everything on your site. Connected off-domain intent signals let the journey — and the score — start before the first visit.