The difference between activity and commercial interest.

What buyer intent data is

Buyer intent data is any observable behavior that suggests an organization is researching, evaluating or preparing to buy in a category. It is inherently probabilistic. No single event proves a purchase is coming, and treating one as though it does is the most common way intent programs lose credibility with sales.

• ✦It is about accounts, not individuals — buying is a committee activity.
• ✦It is a pattern across time, not an event.
• ✦It decays. A signal from last quarter is history, not intent.

First-party and third-party signals are not interchangeable

First-party intent is behavior on property you control: your site, your emails, your product, your events. Third-party intent is behavior observed elsewhere and sold to you — category research on publisher networks, review sites, and similar. They differ in precision, freshness and defensibility, and they fail in different ways.

• ✦First-party is precise and attributable, but only covers accounts that already found you.
• ✦Third-party widens the aperture to in-market accounts that have never visited, at the cost of certainty about who and when.
• ✦First-party signal is yours; third-party signal is, by construction, also being sold to your competitors.
• ✦Only first-party signal can be joined confidently to a specific contact and conversation.

Fit, intent and timing are three separate axes

The most common scoring mistake is collapsing them into one number. An account that is a perfect fit but reading nothing is a nurture target. An account showing heavy activity but structurally wrong for your product is a distraction dressed as a lead. Keeping the axes separate makes the resulting decision explainable to the rep who has to act on it.

• ✦Fit — firmographics, technology, segment. Slow-moving and largely knowable in advance.
• ✦Intent — behavior suggesting active research. Fast-moving and decays.
• ✦Timing — contract cycles, funding, hiring, leadership change. Episodic.
• ✦Relationship — existing contacts, prior conversations, churn history. Often the strongest predictor and the most frequently ignored.

Building a score people will actually use

A score is a communication device as much as a calculation. If a rep cannot see why an account scored highly, they will ignore the score and work their own list — which is a rational response to an unexplained number. Weight the inputs deliberately, show the reasoning, and validate against outcomes rather than against intuition.

• ✦Weight by page and action, not by hit count: the security and pricing pages are not the blog.
• ✦Reward multiple distinct people from one account over one person visiting repeatedly.
• ✦Apply recency decay so the list reflects this week rather than the year.
• ✦Expose the reasoning with the score, so disagreement is informed.
• ✦Backtest against closed-won and closed-lost before trusting the thresholds.

Where first-party signal is captured

Operationalising it without burning the list

Intent programs fail at the handoff more often than at the model. A score that produces an unqualified task queue trains reps to ignore it. What works is routing intent into the same prioritization surface the team already uses, with the context attached and the first action prepared — and holding the outreach to a standard that acknowledges you are responding to research the account did not tell you about.

• ✦Route intent into the existing daily priority list, not a separate dashboard.
• ✦Attach the evidence so the first message can be relevant rather than generic.
• ✦Do not reference surveillance-flavoured detail in outreach; lead with the problem, not with what you watched.
• ✦Measure the program on meetings and pipeline created, not on signals detected.

Turning a ranked account into prepared workOwning the model across teams