What Is Intent Data? Types, Limits, and How Teams Actually Use It

Search for a definition of intent data and you will get eight answers from eight vendors, and almost all of them describe the same thing: accounts researching topics across the internet. That definition is not wrong. It is narrow, and the narrowness is expensive.

If intent data only means research behavior that a vendor observed, the category can tell you an account looks warm and very little else. It cannot tell you who to contact, what changed inside the business, or what to say. Those are the three things that decide whether outreach gets a reply. This page defines the category by where the evidence comes from, and how much weight each kind can carry.

What Is Intent Data?

Intent data is any evidence that a company's situation has changed in a way that makes your product more relevant now than it was last quarter. It comes from two broad places: behavior that somebody observed, and changes a company published about itself.

Both are intent data. Only one of them is something you can verify yourself, and that difference shapes everything you can do with it.

Most buying guides split the category into first-party and third-party data. That split tells you who collected the data, not what it is worth once you sit down to write a message.

Why the Standard Definition of Intent Data Is Too Narrow

The standard definition is too narrow because it describes one collection method rather than the whole category. Topic-consumption data is a source of intent data, not the definition of it. The practical cost is that teams conclude intent data is something you buy, so everything a company publishes about itself for free gets filed under research and never reaches the scoring model or the sequence.

The cost of getting this wrong has gone up. In a survey of 632 B2B buyers conducted in August and September 2024, Gartner found that 73% of B2B buyers actively avoid suppliers who send irrelevant outreach, and that 61% prefer an overall rep-free buying experience. Irrelevant outreach is not a neutral event that gets ignored. It removes you from consideration.

The Four Kinds of Intent Data, Ranked by How Directly You Can Observe Them

Intent data sorts into four types, and the useful axis is not who collected it. It is how directly you can observe the underlying evidence. The more directly you can see it, the more you are allowed to do with it.

1. Declared Signals: The Company Said It Itself

Declared signals are public statements a company makes about its own plans. Job postings, pricing page changes, new product and location pages, funding and leadership announcements, procurement notices, conference talks, regulatory filings, and executive posts all sit here.

These are the most verifiable signals available, and most of them are free. The SEC's EDGAR full-text search covers the full text of electronic filings since 2001, which means a company's own description of its priorities and risks is searchable by anyone.

Declared signals are the only category that reliably tells you what changed, which makes them the only category you can build a message around. Their limit is that a stated priority is not an approved budget. A company hiring for a role has decided the problem matters, not that it will buy anything from you.

2. First-Party Behavior: Activity on Your Own Properties

First-party intent is what people do on assets you own: pricing page visits, repeat sessions, demo requests, content downloads, email and ad engagement. It carries the strongest correlation with purchase, because the person chose to engage with you specifically.

Its limit is coverage. First-party data only sees buyers who already found you, and says nothing about the accounts that should know you exist and do not.

3. Reported Third-Party Intent: A Vendor Tells You Research Is Elevated

Reported third-party intent is a score or surge alert from a data provider, assembled from publisher co-ops, review-site activity, or advertising bidstream exhaust. It is the thing most people mean when they say they bought intent data.

Its strength is breadth: it can flag accounts nobody on your team was watching. Its weaknesses are consistent. The score usually resolves to an account rather than a person, you cannot see the underlying event, and you cannot reference it in a message without sounding like you are surveilling someone.

Data sourced from the advertising ecosystem also carries compliance exposure that published signals do not. In March 2024 the Court of Justice of the European Union ruled that the consent string underpinning the ad-tech industry's Transparency and Consent Framework constitutes personal data under the GDPR. If your provider cannot explain where a signal came from, that is worth asking about before it reaches a sequence.

4. Inferred Scores: A Model's Opinion

Inferred or predictive intent is a model's estimate of buying likelihood, built from firmographics, past deal patterns, and whatever signals the vendor already holds. It is useful for sorting a long list into an order. It is not evidence, and it should never be an input to a message.

This order is not a ranking of value, and each type earns its place. It ranks what each type licenses you to do, and the common mistake is borrowing the authority of a declared signal for a score that has not earned it.

How Do You Tell a Buying Window From Background Noise?

A single signal is rarely a buying window. A buying window is usually two or more independent signals pointing at the same business priority within a short period.

One job posting is a data point. A company that posts two lifecycle marketing roles and a revenue operations manager in the same month, publishes a new pricing page, and announces a VP of Customer Success is telling you something much more specific. Retention, expansion, and revenue process are moving up the priority list, and somebody has been given budget and headcount to fix it.

That combination changes four decisions at once: which accounts go to the top of the list, which persona you contact, which problem the first message addresses, and which outcome you lead with. Signals are also perishable. A hiring cluster that made sense in March is a weak opener in September.

Two related reads: how to combine signals into a real buying window and how long a buying signal stays useful.

The Test That Keeps Signal-Based Outreach Honest

Before you reference a signal in outreach, apply one test: could you show the buyer exactly where you saw it?

If the answer is a link to their job board, their pricing page, their filing, or their own post, the signal is fair to mention and the message gets stronger for it. If the answer is that a vendor scored their domain, you can use it to prioritize your day, but you cannot reference it. Nobody wants to hear that software noticed them.

The test also sets a limit on interpretation. A signal tells you what a company did, not why. Present your reading as an inference the buyer can correct, not as a fact you have discovered about their business.

How to Turn Intent Data Into an Actual Message

The gap between having intent data and getting replies is almost always the message. A signal earns its value at the moment it changes a sentence.

Here is outreach that technically used a signal:

"Hi Sarah, I noticed your company is growing fast. Would love to connect and share how we help teams like yours scale."

Growing fast is not a signal. It is a guess dressed as one, and it applies to half the market. Compare a message built on a declared signal:

"Hi Sarah, I saw you're hiring two lifecycle marketing roles and a RevOps manager this quarter. That combination usually means retention and revenue process are becoming bigger priorities than net-new acquisition. If that's roughly right, the part most teams underestimate is how much manual reporting the new hires inherit in month one. Happy to share what we've seen work, or to be told I've read it wrong."

The second message works because every claim in it is checkable, the interpretation is offered rather than asserted, and it gives Sarah an easy way to correct it.

For the full workflow, see how to turn a buying intent signal into outreach that gets replies.

Where AI Fits in an Intent Data Workflow

The bottleneck in signal-based outbound is not finding signals. It is reading enough of them, quickly enough, to act while the window is open.

A team can do this by hand. Somebody checks job boards, pricing pages, filings, review sites, and executive posts across a target list, works out which changes matter, researches the account, and writes something specific. That work is real, and it does not scale past a few dozen accounts a week.

Alsona is built for that gap. It monitors 31 intent signals across hiring, funding, technology, advertising, reviews, social, and company activity, scores accounts on those unstructured signals alongside ICP fit, researches the account, and turns the context into individualized LinkedIn and email messages, with follow-ups and replies handled in one inbox.

None of that removes the salesperson. Judgment about whether a signal means what it appears to mean stays with the person. What changes is how many accounts can get that judgment applied to them in a week.

The Takeaway

Intent data is broader than the definition the category sells, and the declared signals it leaves out are the ones that tell you what to say. Use all four types for what each is good at: inferred scores to sort, third-party intent to widen coverage, first-party behavior to catch people already looking, and declared signals to decide the message. The teams that get the most out of intent data are rarely the ones buying the most of it.

Frequently Asked Questions

What is intent data in simple terms?

Intent data is evidence that a company may be moving toward a purchase. It comes either from behavior someone observed, such as research activity or visits to your website, or from changes the company published itself, such as a job posting or a funding announcement. It is a timing and context signal, not a guarantee of interest.

What is an example of intent data?

A company posting three revenue operations roles in one month is intent data, because it suggests revenue process has become a funded priority. Other examples include a new pricing page, a competitor complaint in a public review, a leadership hire, a procurement notice, and a spike in research activity on a topic reported by a data provider.

What is the difference between first-party and third-party intent data?

First-party intent data is behavior on properties you own, so it is accurate but only covers buyers who already found you. Third-party intent data is reported by an outside provider and covers accounts you are not yet in contact with, but it is usually account-level and you cannot see the underlying event. Most teams need both. This comparison goes through the trade-offs in detail.

How accurate is intent data?

Accuracy varies sharply by type. Declared signals such as job postings and filings are verifiable, because you can open the source and read it. Reported third-party scores are probabilistic and depend on the provider's collection method and how they match activity to a company. Treat a score as a reason to look closer, not as a finding.

Is intent data GDPR compliant?

It depends entirely on how the data was collected. Signals a company published about itself raise few issues, since the information is public and corporate rather than personal. Data derived from the advertising ecosystem is more exposed, and the Court of Justice of the European Union has already ruled that the ad-tech consent string is personal data under the GDPR. Ask any provider to document their source, and treat this as general information rather than legal advice.

Build Outbound Around Signals You Can Point To

To see this on a real piece of content, Alsona's free podcast signal extractor pulls the intent signals out of any podcast, webinar, or video and drafts an opener from them. It is a small version of the same idea: read what somebody actually said, then write to that.

For a wider view of the signals vendor intent data tends to miss, see our guide to unstructured buying signals and how to act on them, or how intent-based lead scoring puts these signals into a prioritized list.