What is intent data, and what are its three kinds?
Intent data comes in three kinds, by who owns the instrument. Across 22 of our tenants, most intent had only one kind behind it, and one kind is rarely a window.
Intent data is evidence that a company is researching a problem you solve, collected from what its people do online. It comes in three kinds, sorted by who owns the instrument: first-party from your own site and product, second-party from a partner's surfaces such as a review site, and third-party bought from a network of publishers. In our own data across 22 tenants, 71% of the accounts that showed a public company signal, such as a relevant hire or a funding round, had no research behavior on the tenant's own surfaces within a month. One kind of evidence is a hint. Two kinds that agree are a window.
What is intent data?
Intent data is any record of research behavior that suggests a company is moving toward a purchase in your category, used to find accounts on the readiness axis before they fill out a form. It can be a pricing page read three times from one office, a surge in a company's reading about a topic across publisher sites, or a cluster of comparisons on a review site.
In The Demand Compass, intent data feeds one axis of two. It says something about readiness to buy, the momentum inside an account right now. It says nothing about whether the account has heard of you. Read more on that axis in Readiness to Buy.
Almost nobody is buying at a given moment: about 5% of B2B buyers are in market at a time, because companies change providers roughly once every five years. [source: John Dawes, Ehrenberg-Bass Institute with the LinkedIn B2B Institute, 2021] The job of intent data is to find that small slice early.
What are the three kinds of intent data?
The three kinds of intent data are first-party, second-party and third-party, and the difference is who owns the instrument that saw the behavior. First-party comes from surfaces you own, second-party from a partner who shares what happened on theirs, and third-party from a vendor that watches many sites you do not own.
| Kind | Source | Blind spot |
|---|---|---|
| First-party | Your site, product, events | Accounts not yet on them |
| Second-party | A partner, such as a review site | All other surfaces |
| Third-party | A publisher network | Who reads, and if they know you |
First-party data is the most precise and often reaches the person. Second-party data catches late comparison shopping. Third-party data sees research earliest, before anyone visits you, but stops at the company.
Third-party intent is what most people mean by the term. One large provider describes its method plainly: it watches content consumption across a co-op of publishers and B2B brands, and flags a company when the last three weeks of reading on a topic rise above that company's own 12-week baseline. (Source: Bombora, Company Surge product page, 2026)
The book cuts the same ground by level instead: industry, company and lead. A funding round or a new revenue leader moves a company. Repeated visits to your pricing and docs are the lead level, and you only see them if you own the instrument. Both views agree that the most specific evidence is the kind you collect yourself.
Is more intent data better?
More intent data is not better when every source reports the same kind of evidence, because readiness is a composite and one kind cannot corroborate itself. Teams already run several intent providers, and the published research shows their top problem is reaching the right people, not finding more signal.
Forrester's survey of intent data buyers found that over 70% use more than one provider and almost half use three or more. The top execution problem was finding the right people, because few providers deliver signals from known contacts. (Source: Brett Kahnke, Forrester, Intent Data Expectations Vs. Reality, 2023)
Our sales calls show the same pattern. In 7 of 60 external calls between August and November 2026, the other side raised intent data. In 6 of those 7 they had already run at least one intent source, now or in a previous role. In 2 of the 7 the problem they named was acting on the signals, not getting more of them.
What does our own data show about intent?
Our own data shows that most accounts with readiness evidence have only one kind of it. Across 22 DemandCompass tenants over the 90 days to November 11, 2026, 31,841 accounts showed at least one readiness signal, and only 18% had both public company-level evidence and behavior on the tenant's own surfaces.
The breakdown, from our warehouse, across all 22 tenants:
The second cut is stricter. Of the 15,165 accounts that showed a public company-level signal, such as a relevant job post, a funding item or a leadership change, 4,386 (28.9%) also showed behavior on the tenant's own surfaces within four weeks. The other 71% had nothing behind the event.
The third cut is about instruments. In the same window, 21 of the 22 tenants had public company signals flowing and 21 had LinkedIn engagement flowing. Only 4 of the 22 had a website-visit sensor feeding the readiness axis. None ran a third-party topic feed through the system.
These are our tenants, B2B companies selling to technical buyers, not a sample of any market.
What do our data and the research show together?
Together they show that the gap is corroboration, not volume: teams buy and compare intent sources, while our data shows most accounts carry only one kind of evidence, and where the market showed a company event, the missing kind was usually the cheapest one, behavior on surfaces the team already owns.
The book's rule: a company event with no behavior behind it is potential, behavior with no event around it is curiosity, and both on the same account within a few weeks is a window. By that rule, most of the readiness evidence our tenants saw was potential, not yet a window.
A recent Forrester piece adds the uncomfortable half. Sixty-eight percent of buyers start with a preferred vendor in mind, and that vendor wins 55% of the time; intent signals are often the effect of that preference, not its cause. [source: Ian Bruce and Kelvin Gee, Forrester, Preference Matters More Than In-Market Intent Alone, 2026] So intent on an account that has never touched your surfaces is often research toward someone else. That is why the Compass refuses to act on intent alone. Intent without awareness is noise.
The 2023 Forrester survey read intent as a success, with over 85% of users reporting benefits. Both can be true. Intent helps prioritize; it cannot tell a rep who already prefers them.
How should a team use intent data?
A team should use intent data as one input into an account-level score that requires a second, independent kind of evidence before an account counts as ready, with every signal dated and decayed, rather than as a list of hot accounts to hand to sales. Four habits make that work.
- Wire your own first-party signals before buying another feed: pricing, docs, product use, event attendance. In our data they were what most company events lacked.
- Roll every signal up to the account. The person reading is rarely the person signing.
- Require two kinds before routing. A topic surge plus three pricing visits from the same company is a window. Either alone is a watch item.
- Let it expire. The decay rule says a signal is evidence with a half-life, and readiness fades in weeks. See the decay rule.
For how the signals add up to one number, see what account scoring is. For what a topic surge leaves out at the company level, see the comparison on automaterevops.ai.
Questions readers ask next
Is intent data the same as a buying signal?
Intent data is one family of buying signals: research behavior. Buying signals also include company events such as funding, hiring and leadership changes, which say something changed inside the account. The Compass reads both on the readiness axis, and the window is where they meet.
Is first-party intent data more accurate than third-party?
It is more specific, because you know which pages and which product events produced it. Third-party data reaches accounts earlier and further, but at the company level only. We found no independent accuracy study with a published method, so treat any vendor's accuracy figure as a claim.
Can a small team use intent data without buying a tool?
Yes. Resolve your own site and product traffic to companies, add public company events such as job posts and leadership changes, and score both at the account. In our own data, 21 of 22 tenants had public company signals flowing, and that needs no intent vendor.
How long does an intent signal stay useful?
Weeks, not quarters. The book puts the window after a strong signal at thirty to sixty days before the account shortlists without you, so a signal older than that should count for little.
Take last month's intent list, from whatever source you have. For each account, check whether anyone from that company also touched your own site, docs, product or events in the same four weeks. Split the list in two. The accounts with both kinds are your sales list this week. The rest stay on watch, and the share that had only one kind is your corroboration gap.
This note draws on the chapter Readiness to Buy of The Demand Compass: The Signal-Based GTM Framework for B2B Marketing Leaders in Complex Sales, by David Moreira and Marcos Stubrin.