Intent Data: What It Is and How to Actually Use It

The three types, ranked by reliability

The ranking is almost exactly inverse to how much money the market spends on each. Third-party intent is the most heavily marketed and the least reliable, while first-party intent — which is free and already happening on your own website — is routinely uninstrumented at companies paying five figures a year for topic data.

What intent data can and cannot tell you

Third-party intent is the type most people mean, so it is worth being precise about what it delivers. It says: someone at this company, whose identity you do not know, consumed content about this topic, at approximately this time, at an intensity above their baseline.

• It cannot tell you who. The signal is account-level. The person researching may be an intern, a consultant, or someone in an unrelated department.
• It cannot tell you why. Research about your category may be for a purchase, a competitive analysis, a board paper, or a university assignment.
• It cannot tell you where in the process they are. Topic consumption looks the same at problem identification and at final vendor selection.
• It can tell you that something changed. That is genuinely valuable, and it is the whole of the value.

Which leads to the single most important framing: intent tells you when, not who. Using it to select accounts you would not otherwise target is the mistake that produces most disappointing implementations.

Why most implementations fail

1. 01
   Buying intent before defining an ICP

   Intent surfaces accounts researching your topic, including many you cannot serve. Without a fit filter in front of it, you have bought a list of poor-fit accounts with the appearance of urgency.

2. 02
   Treating it as a targeting input

   The correct use is to reorder a list you already believed in. Adding accounts because they showed intent inverts the logic and dilutes your targeting.

3. 03
   Responding too slowly

   Signals decay in days. If the data lands in a weekly report that someone reviews on Friday and acts on the following Tuesday, the window has closed. This is the most common operational failure.

4. 04
   Referencing the signal in the message

   Mentioning that you noticed them researching a topic is unsettling and reveals you know less than you imply. Use the signal to decide who and when; write the message about their situation.

5. 05
   Never measuring it

   Almost nobody runs a control group, so almost nobody knows whether their intent spend produced anything. The vendor's dashboard reports signals delivered, which is not an outcome.

How to act on a signal

The play that works is unglamorous and depends entirely on speed.

Step six is underrated. Even when outbound contact fails, an account showing intent is worth putting in front of, and paid retargeting against a small verified-intent audience is one of the more efficient uses of media budget available.

Measuring whether it works

The only honest measurement is a lift test. Split your ICP-fit accounts into two groups matched on segment and size. Run the intent-triggered play against one; run your normal cadence against the other. Compare meeting rate and pipeline created over a full cycle.

Vendors will discourage this, usually by pointing at influenced pipeline in their dashboard. Influenced pipeline counts every deal where an intent signal existed at any point, which in a reasonably sized account universe is most of them. It is not evidence.

• Primary measure: pipeline created per account contacted, test versus control.
• Secondary: meeting rate and time from signal to first contact.
• Guardrail: cost per qualified opportunity including the intent subscription, fully loaded.
• Run it for a full sales cycle. Anything shorter measures noise.

Is it worth buying?

Third-party intent earns its cost in a specific situation: a large, well-defined ICP that you cannot cover with your current team, where prioritisation genuinely changes who gets contacted this week. If you have 400 target accounts and three reps, you do not have a prioritisation problem — you have a coverage plan, and intent data will not add much.

For most companies the honest sequence is to instrument first-party intent properly, add review-site signals if your category has meaningful traffic there, and only then evaluate third-party topic data against a control group. That order costs almost nothing to follow and frequently reveals that the third step is unnecessary. How these signals sit alongside the others is covered in buying signals.