B2B intent data is sold as a list of accounts that are ready to buy. It is not. It is a probabilistic estimate that people at a company have been reading about a topic more than they usually do, and the gap between those two descriptions is where most intent programs quietly die. This guide explains what B2B intent data actually measures, how surge scoring works underneath the dashboard, the four places it breaks, how to test a provider before you sign, and — the part most vendors skip — exactly which play to run when a topic spikes.
The context is worth holding onto. The Ehrenberg-Bass 95:5 rule suggests that at any moment only around 5% of your market is actively buying. Intent data is an attempt to find that 5% earlier than your competitors. Done well it compresses cycles. Done badly it produces an expensive account list that reps ignore by week three.
What B2B intent data actually measures
Every B2B intent data product answers one question: is consumption of a given topic, by people resolvable to a given company, unusually high right now compared with that company’s own baseline?
Three things follow from that definition, and each one changes how you should use the output.
- It is relative, not absolute. A surge means unusual for that account. A 40,000-person enterprise always reads about security. The signal is the deviation, not the volume.
- It is usually account-level, not person-level. You learn that someone at the company is researching. You rarely learn who, which is why enrichment and persona mapping sit downstream of every intent feed.
- It is topical, not vendor-specific. Research into “data enrichment” does not tell you whether they are evaluating you, a competitor, or building in-house.
Teams that internalise those three constraints write far better plays than teams who treat the dashboard as a buying list.
The three types of B2B intent data
Vendors blur the three types of B2B intent data together in their marketing. Keep them separate in your head, because they have wildly different accuracy and cost profiles.
First-party intent
Behaviour on properties you own: pricing page visits, documentation reads, demo requests, product trial activity, and repeat sessions from one domain. This is the most accurate intent you will ever get and the cheapest to collect. It is also the most under-used, because most teams route only form fills and ignore everything else. If you do one thing after reading this article, instrument your own site properly before buying anyone’s panel.
Second-party intent
Behaviour on someone else’s property that relates directly to your category: review site category browsing, comparison page views, and vendor profile visits on marketplaces. This is the highest-value purchased intent, because a person reading a head-to-head comparison is already in an evaluation. Coverage is narrow by design, so treat it as a scalpel rather than a fire hose.
Third-party intent
Aggregated content consumption across a publisher co-operative or a media panel, mapped to topics and resolved to company domains. Providers such as Bombora and 6sense built the category here. Breadth is enormous, precision is lower, and the quality depends almost entirely on the underlying source mix. Ask what that mix is before you sign anything.
How surge scoring works under the hood
The mechanics of B2B intent data are consistent across vendors even when the branding is not. Understanding the four steps tells you exactly where errors enter.
- Collection. Content consumption events are gathered from a publisher network, a media panel, review sites, or a co-operative of participating vendors.
- Resolution. Each event is mapped to a company, usually via IP-to-company matching, hashed email, or a logged-in identifier. This step is the single largest source of error.
- Topic classification. The content is tagged against a taxonomy of thousands of topics. Broad topics generate noise, narrow topics generate coverage gaps.
- Baselining. The account’s current consumption is compared with its own historical average, and a surge score is produced above a threshold.
Note what is absent from that chain. Nothing verifies that the researcher has budget, sits in a relevant function, or has any authority. That verification is your job, and skipping it is why so many surge lists convert like cold lists.
Where B2B intent data breaks
An honest account of how B2B intent data fails will save you a quarter of wasted effort.
| Failure mode | What it looks like | Mitigation |
|---|---|---|
| Bad account resolution | Surges from co-working spaces, ISPs, VPN exit nodes and agencies | Filter non-ICP domains and known shared IP ranges before scoring |
| Remote and hybrid work | Employee research from home IPs never resolves to the employer | Weight third-party intent lower; lean on first-party and second-party |
| Wrong researcher | An intern, a job seeker or a competitor reading your category | Require a persona match before any rep touches the account |
| Topic taxonomy drift | Everything surges because the topic is too broad | Track 5–10 narrow topics, not 60 broad ones |
| Stale delivery | Weekly batch files arrive after the evaluation closed | Insist on daily refresh and measure the delivery lag yourself |
The remote-work problem deserves particular attention. IP-to-company resolution was designed for a world where people worked in offices. Any vendor claiming unchanged match quality since 2019 is either using a very different method or is not measuring carefully.
How to test a B2B intent data provider before you buy
Run this four-week protocol during the trial. It costs nothing and it settles the argument with evidence rather than case studies.
- The hindcast test. Give the vendor a list of accounts that closed in the last two quarters and ask whether those accounts surged before the opportunity was created. If closed-won accounts did not light up, the feed will not predict the next ones.
- The customer test. Feed them your current customer list. If your own customers do not appear as surging on your core topic, resolution is weak.
- The junk test. Check how many surging accounts are agencies, resellers, students, or companies far outside your ICP. Above roughly 20% is a red flag.
- The lag test. Timestamp when a surge appears in the dashboard versus when the underlying behaviour occurred. Anything beyond a week is a reporting tool, not a signal.
The same discipline applies to contact data underneath the intent layer. Our guide to testing B2B data vendors before you buy sets out the sampling method, and enrichment waterfalls and match rates explains what to expect once a surging account needs contacts attached.
The plays: what to run when a topic surges
Here is the gap in almost every B2B intent data article you will read. Everyone explains the data. Almost nobody says what a rep should do on Monday morning. These five plays cover the realistic scenarios.
Play 1: surge plus fit plus known contact
The account is in ICP, is surging on a narrow topic, and you already hold a relevant persona. Route to the owning AE with a 48-hour SLA. The message never mentions the surge. Instead it leads with a point of view on the topic: a benchmark, a common failure mode, or what three similar companies did last quarter. You are trying to be the useful voice arriving at the moment the question is live.
Play 2: surge plus fit, no contact coverage
The most common case, and the one most teams fumble. Do not send anything yet. Enrich first, build a three-to-five person buying group across the likely economic buyer, the likely user and the likely blocker, then run coordinated outreach across email and LinkedIn. A single email to a single unverified contact wastes the signal entirely.
Play 3: surge on a competitor or migration topic
Topics like “alternatives to”, “migration” and “pricing comparison” are the strongest third-party signals available, because nobody reads them idly. Run a displacement play: switching costs, a migration path, and a proof point from a customer who made the same move. Pair it with technographic data to confirm the incumbent before you write a word.
Play 4: surge on an existing customer
A current customer researching your category is either expanding or shopping. Both need a response, and neither belongs with an SDR. Route to customer success within 24 hours with a simple question about what changed, not a pitch. This play alone often pays for the subscription.
Play 5: surge without persona or fit confidence
Do not put a human on it. Add the account to a paid social and display audience, serve topic-relevant content for 21 days, and watch for a second signal: a site visit, a job posting, or a funding event. One weak signal is a guess. Two independent weak signals become a reason to act.
Intent works best when stacked with dated signals
B2B intent data tells you a topic is warm. It does not tell you why now. Stacking it with a signal that carries a hard timestamp converts a probability into a reason to call.
| Intent surge combined with | Why it becomes actionable |
|---|---|
| A funding announcement | Budget now exists for the thing they were researching |
| A relevant executive hire | Someone was hired specifically to solve this |
| A cluster of related job postings | The initiative is staffed, so it is funded |
| A champion joining the account | You have an internal advocate and a live topic |
| Repeat pricing page visits | First-party confirmation of third-party inference |
This stacking logic is the core of modern outbound. Our guide to the 2026 trigger stack shows how the combinations perform against each other in practice.
Routing, thresholds and the noise problem
Three rules keep a B2B intent data program alive past the first quarter.
- Cap the daily queue. Give each rep a maximum of ten surging accounts per day, ranked. An uncapped feed is ignored within two weeks.
- Require two conditions. Surge alone never triggers human outreach. Surge plus ICP fit, or surge plus a second signal, does.
- Expire everything. A surge older than 14 days is history. Close it automatically and stop paying attention to it.
None of this survives a broken foundation. If your account records are duplicated and your emails bounce, an intent feed simply routes reps to bad data faster. See why AI fails without a data layer for the underlying argument.
Measuring whether B2B intent data pays for itself
Most teams measure activity on surging accounts, which proves nothing. Measure lift instead, using a genuine holdout.
- Holdout comparison. Work surging accounts in one segment and a matched non-surging segment in another. Compare meeting rates over eight weeks.
- Precision by topic. Percentage of surging accounts that produced a qualified conversation, broken out per topic. Kill the topics below your baseline.
- Time to first meeting. Whether arriving earlier actually shortens the path, or simply adds touches.
- Cost per sourced opportunity. The full subscription cost divided by opportunities the program genuinely originated.
B2B intent data FAQ
Is B2B intent data accurate enough to build outbound on?
On its own, no. As a prioritisation layer on top of a well-defined ICP, yes. The realistic expectation is that it reorders your existing target list rather than generating a new one. Teams that expect a ready-to-buy list are always disappointed.
How many topics should we track?
Five to ten narrow ones. Every additional broad topic dilutes the signal and inflates the queue. Choose topics a buyer would only research while solving the specific problem you solve, then review precision by topic each quarter.
Should we mention the intent signal in outreach?
Never directly. Telling someone you noticed their company researching a topic is unsettling and often wrong, since you cannot know it was them. Use the signal to choose the subject, then write as though you simply have a point of view worth reading.
Is buying B2B intent data compliant in the EU?
Account-level intent generally sits on firmer ground than person-level tracking, which requires consent under GDPR. Ask any provider for its lawful basis, its source disclosures and its EU coverage caveats in writing. Our outbound compliance guide covers the wider rules.
Treat intent as a ranking layer, not an answer
B2B intent data earns its keep when it changes the order of a queue you were going to work anyway, and when every surge has a named play behind it. Instrument your own site first, buy a narrow third-party feed second, require two conditions before a human touches anything, and measure lift against a holdout. The vendors will keep selling the dashboard. The pipeline comes from the play.
Next, see how to act on the fastest-moving triggers in our guide to sales trigger events, or browse more on B2B data and sales intelligence.
Ready to act on intent instead of admiring it? ZenBee combines buying and hiring signals across 35M+ companies with 700M+ verified contacts and multichannel LinkedIn and email outreach in a single workflow. Request a demo, or start for free.