B2B Intent Data: What It Actually Predicts and Which Providers Are Worth Paying For
Intent data predicts that an account is researching a topic more than usual. That is the whole signal. What the providers cost, and why most programs disappoint.
B2B intent data reliably predicts one thing: an account is consuming more content about a topic than its own historical baseline. It is account-level, refreshes weekly, and is sold to competitors too, so it works as a timing and prioritisation layer rather than a targeting source. Third-party feeds typically cost $15,000 to $150,000 a year.
Key takeaways
- Third-party intent data is account-level and typically refreshes weekly, so it cannot tell you which person at the account is interested.
- Bombora's Company Surge is the upstream co-op source that several other platforms license and resell, so buying two vendors can mean buying the same signal twice.
- Reported annual costs run roughly $30,000 to $60,000 for Bombora, $50,000 to $150,000 for 6sense, $20,000 to $50,000 for G2 Buyer Intent, and from about $15,000 for ZoomInfo Intent as an add-on.
- First-party signal from your own site and product is stronger, cheaper and exclusive, and most companies underinstrument it before buying a third-party feed.
- Intent signals are perishable, so each one needs a predefined action, a named owner and a response window of 24 to 72 hours.
- Never reference the surveillance in outreach; use the signal for timing and write to the underlying problem instead.
Reviewed and updated August 10, 2026
B2B Intent Data: What It Actually Predicts and Which Providers Are Worth Paying For
Intent data predicts one thing reliably: that an account has recently consumed more content about a topic than it usually does. That is the whole signal. Everything else, including which person is interested, whether budget exists, and whether anyone is buying this quarter, is inference layered on top by you or by a vendor's model.
Teams that understand that get real value from intent data. Teams that expect it to hand them buyers get a $40,000 annual contract and a dashboard nobody opens.
First-party versus third-party intent
The two categories behave so differently that treating them as one budget line is the first mistake.
| First-party intent | Third-party intent | |
|---|---|---|
| Source | Your own properties and systems | Publisher co-ops, review sites, bidstream |
| Resolution | Often person-level | Almost always account-level |
| Latency | Real time to hours | Typically weekly refresh |
| Exclusivity | Yours alone | Sold to your competitors too |
| Typical cost | Low, mostly instrumentation | $15,000 to $150,000 per year |
| Predictive strength | High | Moderate, directional |
First-party intent is your pricing page views, docs and changelog traffic, repeat visits, free-trial behaviour, email engagement, and views of your own review-site listing. It is stronger than anything you can buy, because the account is already looking at you. Most companies underinstrument it and then pay for third-party data to compensate. If you are not already identifying and routing your own website traffic, start with website visitor identification, which is cheaper and more actionable than any co-op feed.
Third-party intent tells you an account is researching a category somewhere else on the internet. Useful for timing and prioritisation. Not sufficient for targeting on its own.
How third-party intent is actually produced
Three mechanisms, and knowing which one you are buying tells you what the data can support.
Publisher co-ops. Bombora's Company Surge is the reference implementation: a cooperative of thousands of B2B publishers contributes content-consumption events, which are resolved to companies by IP and account matching, then compared against that company's own historical baseline for the topic. The output is a surge score, not a raw volume, which is why a large enterprise browsing normally does not outrank a small company suddenly reading everything about your category. Critically, Bombora is the upstream source for much of the market. Several platforms license or resell Company Surge inside their own product, so buying two vendors can mean buying the same signal twice.
Review and comparison sites. G2 Buyer Intent reports accounts viewing your category, your profile, and your competitor comparison pages. Narrower coverage than a co-op, and much closer to purchase.
Predictive platforms. 6sense and similar layer their own model on top of licensed co-op data, website de-anonymisation, and CRM history to output a predicted buying stage per account. You are buying the model and the orchestration, not just the signal.
Provider landscape
Intent vendors do not publish list prices. The ranges below are the figures consistently reported by buyers and third-party pricing guides through mid-2026, and real contracts vary widely with account volume, topic count, and bundling. Treat them as a budgeting order of magnitude, and get your own quote.
| Provider | What you get | Reported annual range | Best fit |
|---|---|---|---|
| Bombora | Company Surge topic data, the upstream co-op source | $30,000 to $60,000 standalone | Teams that want raw signal to route into their own systems |
| 6sense | Predicted buying stage, ABM orchestration, visitor ID, licensed intent | $50,000 to $150,000+ | Enterprise ABM with the headcount to operate it |
| G2 Buyer Intent | Accounts viewing your category, profile, and comparisons | $20,000 to $50,000 | Software vendors already active on G2 |
| ZoomInfo Intent | Intent as an add-on beside a large contact database | From around $15,000 as an add-on | Teams already on ZoomInfo who want signal plus contacts in one place |
If you already own a data platform, the add-on route is usually the right first purchase. Compare the underlying data layers in best ZoomInfo alternatives before assuming your current vendor is the cheapest path to signal.
Why intent data disappoints
Every failed intent program we have seen shares the same causes. None of them are data quality.
No activation motion exists. The account list arrives weekly and lands in a spreadsheet. Intent data is perishable: a surge that is three weeks old is history. If there is no sequence ready to fire within days, the signal has no way of becoming pipeline.
Account-level signal, person-level outreach. The feed says Acme is researching your category. It does not say who. You still have to pick contacts, and if you pick the wrong five people you get the same reply rate as an unprioritised list. Intent narrows the account set; it does not write the list.
Everyone gets the same feed. Your competitors buy the same co-op. Being first to a surging account is an operational advantage, not an informational one, and it is won by speed of response.
Surges are noisy. Job seekers, students, analysts, consultants, and your own employees all generate content consumption from a corporate IP. Filter to ICP-fit accounts before you look at scores, never after.
It gets blamed for a bad offer. Prioritisation multiplies whatever your message already does. Multiplying a 0.4 percent reply rate by two is still a bad campaign. Fix the motion first: our cold email reply rate benchmarks show what a healthy baseline looks like, and 4 frameworks that fix broken lead generation covers the diagnosis.
An activation model that works
Intent data pays off when each signal has a predefined action, an owner, and a time limit. Write the table before you sign the contract.
| Signal | Action | Owner | Response window |
|---|---|---|---|
| Third-party topic surge on an ICP-fit account | Add to a targeted sequence with a category-relevant angle | Outbound | 72 hours |
| Person-level website visit on a pricing page | Direct outreach referencing the relevant problem, not the visit | AE or founder | 24 hours |
| Repeat visits from multiple people at one account | Multithread across the buying group | AE | 48 hours |
| Competitor comparison page view on a review site | Competitive displacement angle | AE | 24 hours |
| Surge with no ICP fit | Ignore, and exclude the account from scoring | Nobody | None |
Two rules make this work in practice. Never reference the surveillance itself: "I saw you were researching X" converts badly and reads as creepy. And keep the response window short enough that the signal is still true when the message lands. The channel matters less than the speed, which is why we run intent-triggered accounts through both email and LinkedIn outbound simultaneously rather than sequentially.
Should you buy it yet?
A blunt readiness test. Buy third-party intent data only if all four are true:
- Your unprioritised outbound already converts at a defensible rate.
- You have instrumented first-party signal and are acting on it.
- Someone owns response inside 72 hours, with a sequence already written.
- Your ICP is large enough that prioritisation is a real problem. Under a few hundred target accounts, just contact all of them.
If fewer than three are true, spend the money on list quality and message testing instead. The 7-layer GTM AI stack sets out the order these layers should be built in, and the outbound sales playbook covers the motion that intent data is supposed to accelerate.
Bottom line
Third-party intent data is a timing and prioritisation layer that costs five figures a year and returns nothing without an activation motion behind it. First-party signal is cheaper, sharper, and almost always under-exploited. Build the motion, instrument your own traffic, then buy the co-op feed to widen the funnel.
If you want the motion built before the data spend, get a free campaign plan and we will show you what your current list and message produce first.
Frequently asked questions.
Frequently asked questions- What does B2B intent data actually tell you?
- That an account has consumed more content about a topic recently than its own historical baseline. That is the complete signal. It does not identify which person is interested, whether budget exists, or whether a purchase is planned. Those are inferences layered on top by you or by a vendor's predictive model, and they carry the model's error rate.
- How much does intent data cost?
- Vendors do not publish list prices. Buyers and pricing guides consistently report roughly $30,000 to $60,000 a year for Bombora standalone, $50,000 to $150,000 for 6sense, $20,000 to $50,000 for G2 Buyer Intent, and from about $15,000 for ZoomInfo Intent as an add-on. Real contracts vary with account volume, topic count and bundling.
- What is the difference between first-party and third-party intent data?
- First-party intent comes from your own properties: pricing page views, product usage, email engagement, repeat visits. It is often person-level, real time and exclusive to you. Third-party intent comes from publisher co-ops, review sites and bidstream data, is almost always account-level and weekly, and your competitors buy the same feed.
- Why has intent data not improved our reply rates?
- Usually because prioritisation multiplies an existing motion rather than creating one. If the underlying message and list convert poorly, ranking the same accounts differently changes little. The other common causes are latency, where signals are worked weeks after they surge, and treating account-level signal as if it named a person to contact.
- Should a small B2B company buy intent data?
- Usually not yet. Buy it only when unprioritised outbound already converts, first-party signal is instrumented and acted on, someone owns response within 72 hours with a sequence written, and your target account list is large enough that prioritisation is a genuine problem. Under a few hundred target accounts, contact all of them instead.
About the author.
Fernando Cao is CEO at RevenueFlow, which builds and operates outbound revenue engines for B2B companies. Previously at Accenture Strategy. Studied at University of Bath.
Fernando Cao · CEO
Connect on LinkedIn →Explore more.
Ready to scale your outreach?
We build GTM engines that book real meetings. See the receipts.
Related articles.
Website Visitor Identification: Turning Anonymous Traffic Into Named Accounts
Most B2B site visitors never fill in a form. Company-level and person-level identification recover some as named accounts, with very different legal exposure.
How to Build an ICP That Actually Changes Your Target List (With TAM Math)
Most ICPs never change the target list. Here is the count test, the five inputs, a worked TAM, SAM and SOM example with full arithmetic, and an eight-field worksheet.