What Intent Data Is, and Whether It Works

Intent data is one of the most oversold categories in B2B marketing, and also one of the genuinely useful ones, which is a confusing combination. Sold well, it promises to tell a sales team exactly which companies are about to buy. Used well, it does something much narrower and more honest. It helps a team decide who to call first. The gap between those two descriptions is where most of the disappointment lives.

The two kinds of intent

There are two main flavors, and they are not interchangeable. First-party intent is behavior a company observes on its own properties. Someone visits the pricing page three times in a week, downloads a comparison guide, opens four emails, or returns to the product after a trial lapsed. This is direct evidence. The company owns it, sees it in real time, and can trust what it means because it watched it happen.

Third-party intent is inferred from behavior happening somewhere else. Providers track research activity across networks of publisher sites and content properties, then report when a company appears to be reading more than usual about a given topic. The best known model, Bombora's Company Surge, works through a cooperative of B2B publishers and brand sites that contribute consumption data, scored against each company's own historical baseline so a spike registers as a surge. The signal is real, but it is a step removed from the buyer and from the brand doing the buying.

What the signals actually tell you

The crucial limitation of most third-party intent is that it reports at the company level, not the person level. A surge score says an account showed elevated research on a topic. It does not say who at that company was reading, whether they have any budget, or whether the research was for buying, for a competitive review, or for a college student's term paper. Topic surges are also broad. Interest in a category is not the same as interest in a specific product, and certainly not the same as readiness to talk to a salesperson.

• First-party intent: who is engaging with the brand right now, high fidelity, narrow coverage
• Third-party intent: which companies may be researching a topic elsewhere, lower fidelity, wide coverage
• Company-level signal: tells you an account is active, not which person or why
• Topic surge: signals category interest, not product interest or buying readiness

The honest take on accuracy

Intent-data accuracy is contested, and any single number a vendor quotes deserves skepticism. First-party signals are about as accurate as data gets in this space, because they are observed rather than inferred. Third-party signals are directionally useful and meaningfully noisier. Providers attribute behavior to companies through identity graphs that map IP addresses and devices to organizations, and that mapping is imperfect, especially with remote work scattering employees across home networks. The result is real signal mixed with real noise, in proportions that vary by provider and by how niche the topic is.

It is worth treating every accuracy figure as self-reported. Vendors measure their own data against their own definitions of a hit, and there is no neutral referee. The reasonable assumption is that third-party intent surfaces some genuine in-market accounts earlier than a team would have found them otherwise, alongside a meaningful share of false positives that waste reps' time if treated as gospel.

Coverage adds another wrinkle that vendors rarely volunteer. A co-op only sees research that happens on the sites it can observe, which skews toward large, well-trafficked publisher properties. A buyer who does most of their research inside a private community, a vendor's own documentation, or increasingly an AI assistant that summarizes the web for them may leave almost no trail in the co-op at all. As more early research moves into channels that intent networks cannot see, the blind spots grow, and a quiet account is no longer safe to assume is an inactive one. Intent data is a window onto part of the market, not a mirror of all of it.

Where it earns its keep

Intent data delivers the most value as a prioritization layer rather than a decision engine. The right question to ask of it is not what will this account buy but which of these thousand accounts deserves attention this week. Used that way, the noise becomes tolerable, because the cost of a false positive is one wasted outreach attempt rather than a strategy built on a bad assumption.

• Account discovery: surfacing companies researching a category before they hit the website
• Outbound prioritization: ranking a target list so reps work the warmest accounts first
• ABM targeting: timing campaigns to when an account shows elevated activity
• Renewal and expansion: catching existing accounts researching competitors

The strongest results tend to come from combining the two flavors rather than betting on one. Third-party intent widens the net and catches accounts the brand has never touched. First-party intent confirms which of those accounts actually showed up and engaged. An account that surges on a third-party topic and then visits the pricing page is a far better bet than either signal alone.

Layering also dampens the weakness of each source. Third-party data is broad but noisy, so on its own it produces long lists with a low hit rate. First-party data is precise but narrow, so on its own it only ever sees the accounts already in motion on owned properties. Stacked together, the third-party signal nominates accounts to watch and the first-party signal scores how serious they are once they engage. The accounts where both light up are the ones worth a rep's time this week, and the gap between a single-signal list and a two-signal list is usually the difference between intent data that reps trust and intent data they quietly ignore.

How to use it without getting burned

The teams that get value from intent data share a few habits. They feed it into scoring and routing rather than handing raw surge lists to reps and calling them leads. They pair every third-party signal with a first-party check before treating an account as hot. They resist the urge to read a topic surge as a purchase forecast. And they measure intent-sourced outreach against a control, so they can tell whether the data is actually lifting conversion or just rearranging the same pipeline.

Does intent data work? As a way to know what a buyer will do, no, and any pitch that promises that is overselling. As a way to spend limited sales attention where it has the best odds of mattering, yes, with the caveat that the signal needs confirming before anyone acts on it as if it were certain. The honest version of intent data is less exciting than the sales deck and considerably more useful.