B2B Intent Data 2026: What It Is, Types and Real Prices

B2B intent data is behavioral evidence that a company is researching a problem, a product category or a competitor right now. It comes from three places: your own channels (first party), a partner site such as G2 (second party) or a data cooperative such as Bombora (third party). Platforms cost a median $25,000 to about $69,000 a year according to Vendr, so the real question is which signals are worth that money.

This guide answers that with numbers: the nine types of intent data, how each is collected, how much each tells you, what the vendors actually charge in 2026, and a simple test for whether bought intent will pay off for you.

One disclosure first: Revexa builds Leo, a buying-signal agent that acts on public signals. Read with that in mind. Every vendor fact below links to the vendor's own page or to a named price database.

What is intent data?

Intent data is any recorded behavior that suggests a buyer is moving toward a purchase. Reading three articles about SOC 2 audits, comparing two CRMs on G2, visiting your pricing page twice in a week, or liking a competitor's product post on LinkedIn are all intent signals.

The vendors define it in similar words:

• Bombora: intent data "reveals when buyers (accounts and buying-group members) are actively researching topics across the internet, based on the content they consume."
• 6sense: "the digital signals that indicate when a company is actively researching a solution like yours."

Most guides skip the honest caveat: intent data is a probability, not a certainty.

That caveat is the key to using it well. Intent data does not tell you who will buy. It tells you who is more likely to be in market than yesterday, and it needs to be combined with fit (is this company a customer type you win?) and a person (who inside the company is actually looking?).

What is B2B intent data, specifically?

B2B intent data is intent data resolved to a company rather than a consumer. Most of it is account level: a provider sees that someone at a given company, identified by IP address or a known domain, consumed content on a topic above that company's normal baseline. Only a few sources tell you which person is researching, and in Europe those sources are narrow (more on that below).

What is an intent database?

An intent database is the store a provider builds from those signals: companies, topics, scores and timestamps, refreshed daily or weekly. When a vendor says it tracks "25,300+ intent topics" (Bombora), that is the size of its taxonomy, not the number of buyers in it. Ask any vendor how many companies in your ICP and your countries show up in a typical week. That number is what you are paying for.

First, second and third party intent data

First party data is strong because it is specific: someone looked at your product. Third party data is broad because it is general: someone at the company read about the category. Second party sits in between, which is why it is often the best value per signal.

The 9 types of intent data and how each is collected

Most guides stop at "first vs third party". That hides the differences that matter for accuracy and cost. Here is the full map, grouped by how the signal is actually captured.

A few numbers help size these sources:

• G2 says "over 200 million software buyers research across G2, Capterra, Software Advice, and GetApp each year" (G2).
• TechTarget reports "58M+ permissioned members across 220 media properties" (TechTarget).
• Bombora tracks "25,300+ intent topics" (Bombora) across "thousands of B2B media sites" and says 86% of the signals in its cooperative are shared only with Bombora (Bombora).
• 6sense tracks custom keywords "in over forty languages" and blends in Bombora, G2, TrustRadius and TechTarget data (6sense).

What is bidstream intent data?

Bidstream data is collected from the bid requests that ad exchanges broadcast when a page loads an ad. Those requests carry the page URL and an IP address, so a reseller can infer that "someone at Company X read an article about Y". It is cheap and huge, which is why it shows up in low-priced intent feeds. It is also the noisiest type: the reader never chose to share anything with the data provider, shared office and VPN IPs blur the company match, and the consent chain behind programmatic ads is contested in Europe. If a vendor will not say what share of its data is bidstream, assume it is a lot.

What is predictive intent data?

Predictive intent data is not a separate source. It is a model that combines the sources above, usually with firmographics and your past wins, and outputs a score or buying stage per account. 6sense is the best known example. The appeal is prioritization. The risk is that you cannot see why an account scored high, which makes it hard to write a relevant first message. Use predictive scores to rank accounts, then look for a concrete, visible signal before you reach out.

What is real-time intent data?

"Real time" usually means daily or weekly batches for third party data. Truly real-time signals are the ones you capture yourself (a pricing page visit, a reply) and public events you monitor (a new job post, a comment on a competitor's launch post). If speed matters for your motion, the first party and public signals are where you get it.

How much each signal is worth: the signal value ladder

Not all signals are equal. We rank them by three questions. We call it the WHO, WHY, WHEN test:

• WHO: do you know the person, or only the company?
• WHY: does the signal point at your problem, or only a broad topic?
• WHEN: is it from this week, or from a monthly aggregate?

A signal that answers all three is worth acting on today. A signal that answers one is only worth a place in a score.

The pattern: signals that name a person and point at your problem beat anonymous topic consumption every time. That is why the free signals you already own often outperform a five-figure feed. More examples in our guide to 30 B2B buying signals.

Intent data pricing in 2026

List prices are rarely public. The table uses the vendor's own pricing page where one exists, and Vendr's database of real purchases otherwise (Vendr figures are medians of deals Vendr has seen, updated February 2026).

Two things stand out. First, the gap is enormous: from free to six figures for what vendors all call "intent data". Second, the middle of the market is disappearing into bigger platforms. Warmly announced it is joining HubSpot (June 30, 2026), Zoom agreed to acquire Common Room (July 2, 2026), and Pocus joined Apollo (March 19, 2026). Signals are becoming a feature of the CRM or sales platform you already pay for.

How accurate is intent data?

No independent study measures false positive rates across providers, and anyone who quotes a single "accuracy" number for intent data is quoting marketing. What the primary sources do say:

• Coverage is partial. RB2B, one of the few vendors to publish it, states 15 to 20% contact-level coverage for visitor identification, and 35 to 45% on its Pro+ plan (RB2B).
• Most accounts are not in market. 6sense itself says "only 5 to 10% of accounts in your ICP are ready to buy right now" (6sense). A feed that flags 40% of your list is flagging noise.
• It is a probability. Intent data shows likelihood, not certainty, so combine it with fit and engagement data.
• Account level hides the buyer. A topic surge at a 2,000-person company might be one intern writing a report.

Timing is the reason people pay anyway. In the 6sense 2025 Buyer Experience Report (nearly 4,000 respondents), buyers made first contact with sellers at about 61% of their journey, buyers started 79% of those first conversations, and the first vendor buyers spoke with won roughly 8 of 10 deals. If most of the decision happens before anyone fills in a form, signals are the only way to see it.

Is intent data worth paying for?

It pays off when three things are true at once:

1. Your addressable market is large. If you sell to 300 named companies, you can watch all 300 directly. A feed adds little.
2. Your deal size carries the cost. Use the break-even test below.
3. You have capacity to act within days. A signal that sits in a dashboard for three weeks is worth nothing.

The break-even test. Divide the annual cost of the data (plus the hours to work it) by your gross profit per deal. That is the number of extra deals the data must create in a year.

If you sell $5,000 deals, a mid-market intent platform needs five to six additional wins a year before it breaks even, before counting the time to act on it. If you sell $50,000 deals into a global market, one extra win pays for it. That is the honest answer: bought third party intent is built for enterprise deal sizes and large account lists. For a company with 10 to 100 employees, the math usually fails on all three conditions at once.

Free intent data you already own

Before you buy, exhaust what you have. These cost nothing beyond the time to look:

1. Your CRM. Leads who asked for a proposal and went quiet are documented interest. See how to find dormant leads and why so many leads never get followed up.
2. Your website. Company level visitor identification has free tiers (Leadfeeder Lite, RB2B free).
3. Your email and product data. A former lead who clicks again, a trial user who invites a colleague.
4. Public events. Job posts, funding, new leaders and tech changes are free to monitor.
5. Public LinkedIn engagement. People who comment on or react to your competitors' and category influencers' posts are showing interest in the problem, by name, this week.

The fifth is the most underused signal in Europe. Person level website identification does not work on EU traffic (RB2B's contact level ID is "US Only"), so public engagement is one of the few signals that tells you who, why and when in the same event.

How to use intent data: from signal to message in 5 steps

1. Define the signals that matter for your business. Write down the three events that preceded your last ten deals.
2. Filter for fit first. A perfect signal at a company that cannot buy from you is noise. Check the business model, not just the job title.
3. Find the person. Account level signals need a named human: the one who owns the problem you solve.
4. Reach out on the context, not the surveillance. "I saw you on our pricing page" is unsettling. A relevant message about the problem the signal points to is not. See how to write the first line.
5. Measure replies and meetings per signal type. Keep the signals that book meetings, drop the rest.

Why the discipline matters: Instantly's 2026 cold email benchmark puts the average reply rate at 3.43%, while the top 10% of senders reach 10.7% or more. Its advice for closing that gap is to blend campaign data with buyer signals such as hiring, funding, launches and site visits, "to reach prospects when they actually care".

Intent data and GDPR in Europe

• First party data about people who engaged with you can usually rest on legitimate interest, with a documented balancing test and an easy opt-out.
• Bought data puts the burden on you to check the vendor's collection and legal basis. Responsibility follows the data.
• Person level web identification is effectively a US product today.
• Scraping contact details is risky. France's CNIL fined Kaspr €240,000 for collecting LinkedIn contact data users had chosen to hide (CNIL).

For B2B outreach specifically, read GDPR and B2B email.

Where Leo fits

Leo is Revexa's AI GTM agent, built as a buying-signal agent rather than an intent data feed. Instead of selling you a dashboard of surging accounts, Leo acts on the signals that name a person: who engages with your competitors' and influencers' LinkedIn posts, plus hiring, funding and leadership changes. It then:

1. Scores each company against an ICP it builds from your website, judging business model, not only title.
2. Finds the right person at that company.
3. Writes the message in your voice, on LinkedIn and email.
4. Waits for your approval. Review mode is the default.

In an internal measurement, the ICP Leo built from our own site matched 19 of 25 of our founder's own account judgments on day one. Leo costs $99 or €89 a month with 2 LinkedIn profiles and 1 mailbox, the first 7 days free and no lock-in. Start Leo and see which companies show buying signals this week.

Keep reading in the cluster: B2B buying signals · What is AI GTM? · Lead generation cost 2026 · How much are leads worth?

Frequently asked questions

What is B2B intent data in simple terms?

It is evidence that a company is researching something you sell: reading about the topic, comparing vendors on review sites, visiting your site or engaging with competitors. It comes as first party (your own channels), second party (a partner like G2) or third party (a data co-op like Bombora).

Is intent data worth paying for?

Only if your market is large, your deal size is high and you can act within days. Divide the annual cost by gross profit per deal to get the extra wins needed. At a Bombora median of $25,000 a year and $5,000 deals, that is five extra deals a year before break-even.

How much does intent data cost?

From free to six figures. Visitor identification starts free or from €79 (Leadfeeder) or $79 (RB2B) a month. Vendr's February 2026 medians are $25,000 a year for Bombora, $62,220 for 6sense and $68,591 for Demandbase.

How accurate is intent data?

There is no independent accuracy benchmark, and intent data is a probability, not a certainty. RB2B states 15 to 20% contact-level coverage, and 6sense says only 5 to 10% of ICP accounts are in market at any time, so treat intent as a ranking input, not proof.

What is the difference between intent data and buying signals?

Intent data usually means aggregated, account level research activity sold by data vendors. Buying signals is the broader term: any event that suggests timing has changed, including hiring, funding, a new decision maker or public engagement with a competitor. Many of the strongest buying signals are free.

Is there free intent data?

Yes. Your CRM history, website visitors (Leadfeeder and RB2B have free tiers), email engagement, job posts, funding news and public LinkedIn engagement all cost nothing to monitor.