What are intent signals?

Open your analytics, your ad platform, and your CRM on the same Monday and you will count thousands of events. Nearly all of them say nothing about a purchase. Someone scrolled, someone opened an email on a phone, a contractor skimmed the docs. The few that do matter are intent signals. Three things separate them from the rest: who did it, where it happened, and how close it sits to a decision. A security review requested by a director at a target account says more than a hundred anonymous blog visits. The label covers everything from a replayed demo(opens in a new tab) to a topic surge on a publisher network. The working skill is sorting signals by kind before anyone scores or routes them.

A signal is an event with an owner and a date

An event becomes a signal when you can attach it to an account or a person, put a date on it, and plausibly place it on the path to buying. Reading a case study counts. Opening a newsletter is weak evidence on its own, and a visit to the integrations page last week says more than the same visit last spring.

Signals also differ in weight. A post impression sits far from a purchase and a pricing page return visit sits close. A system that counts every event equally ends up ranking the accounts that make the most noise, so weighting by distance from a decision is worth the effort. Clean the input as well. Your own employees, existing customers, vendors, and job seekers all produce events that look like interest, and they should be excluded before anything is counted.

Three questions sort most signals

The first question is source. A signal you observed on your own site, ads, and CRM can be traced to a page and a timestamp. A signal from a partner property, such as a review site that shares which companies compared you with a competitor, keeps its source attached. It sits one step off your domain. A signal from a vendor's panel arrives as a topic score, with no page or timestamp behind it that you can check. The second question is level: whether the event belongs to a named person or only to a company. The third is distance from purchase, early research versus a step you would expect near a decision.

Hiring waves, funding rounds, and leadership changes usually travel under a different name. Vendors call them trigger or buying signals, since they describe a company's situation and say nothing about what anyone is reading. The guide to intent signals gone wrong(opens in a new tab) covers what happens when a team skips this sorting and reacts to every event the same way.

Signals, intent data, and intent scores

Signals are the events. A collection of them, bought or observed, is intent data(opens in a new tab), and a score is a number a platform derives from the events. The three terms get used as if they meant the same thing, which hides where a result came from. When someone says an account "has high intent," ask which events produced that. A good answer names them. Working with intent data(opens in a new tab) covers datasets and scores in more detail.

Account-level versus contact-level signals

An account-level signal tells you a company is in motion. It can add the account to an ad audience or raise its place in a ranking, and it cannot tell you who inside the company is doing the reading. A contact-level signal attaches the same event to a named person, and that gives a rep someone to call and a reason to give. Contact-level signals still work inside an account-based program, where the account decides how much budget a signal is worth. It also depends on confident visitor identification(opens in a new tab). A name guessed from a company-level match turns a signal into a rumor, and reps soon learn which alerts to ignore.

How collection has changed

Signals collected through cookies behave differently by browser. Chrome kept third-party cookies after Google dropped its plan to remove them, and Google retired the Privacy Sandbox replacement technologies in October 2025. Safari blocks third-party cookies by default. First-party signals on your own domain are unaffected by any of this. Panel and bidstream data lean harder on third-party tracking, and European regulators have questioned the consent framework bidstream data relies on. Some buyer research also now happens in AI chat tools, which sit outside your analytics, so a share of real research will leave no signal at all.

What an intent signal can't tell you

One signal is evidence of interest, and interest is a long way from a purchase. A signal also says nothing about fit: a strong signal from an account outside your ideal customer profile is still the wrong account. Timing is the last gap, since the event records that something happened and not what the person was trying to decide.

A signal is enough to put an account in front of an ad. A phone call needs more, usually a second signal from a named person at an account that fits. How fast to act is a separate decision from whether to act, and the state of ad signals(opens in a new tab) covers the advertising side of that choice.