What is B2B intent data?
B2B intent data is evidence that a company is researching a topic related to what you sell. It usually comes from content consumption tracked across a network of B2B publishers (offsite or third-party intent, such as Bombora), from review-site activity (such as G2), or from your own website and campaigns (first-party intent). A buying signal is the broader idea: any behaviour that changes the odds an account will buy. That includes the pages a buying committee reads on your site, how often they come back, and how much attention they give your ads.
Good intent data is one of the most powerful early signals in B2B. It shows you accounts researching your space long before they ever reach your site. It earns its keep when you combine it with your own first-party behaviour and someone has a clear job for acting on it. Most teams buy the data and skip that second part.
I’ve worked in adtech since 2011, and our analytics and media business lives on this stuff every day. This hub pulls together everything I’ve written on intent, signals and the gap between buying data and getting pipeline from it.
Where to start
B2B Intent Data: Why Most Teams Use It Wrong: the operating-model problem behind most intent failures.
Segment vs Signal: How to Actually Run B2B Programmatic: why licensed segments disappoint, and what to use instead.
What Are High Value Actions (HVAs)?: the first-party signals that tend to come before pipeline.
B2B Account Scoring: A Model That Beats Fit and Intent Alone: how to put the signals together into a score.
Why most teams don’t get full value from intent data
There’s a story I’ve watched play out inside dozens of B2B teams. Marketing buys intent data. The dashboards go live and accounts start “surging”. Then the surges get exported as a static list or fired straight at SDRs as alerts, and nothing else changes. Six months later pipeline impact is unclear, sales says the data is noisy, and marketing says sales isn’t following up. The data was fine. Nobody built the plan for using it. I set it out in full in B2B Intent Data: Why Most Teams Use It Wrong.
The data is genuinely powerful. The problem is almost always that the organisation invested in collecting signals without building the model that turns them into action. The common mistakes:
A surge is not a sales alert. Treat a surging account as a hot lead, have an SDR call too early, and your domain gets mentally tagged as “too salesy”. In a buying process that Gartner’s research suggests can run 12–18 months for enterprise deals, burning that early relationship is expensive (cited in the same post).
Most of the market isn’t buying right now. The LinkedIn B2B Institute’s 95-5 rule, based on Professor John Dawes’s work at the Ehrenberg-Bass Institute, says only around 5% of a category’s buyers are in market at any given time (Segment vs Signal). Intent is brilliant at spotting accounts moving into that 5% early, but only if you’re already reaching them with always-on, account-level advertising so you’re known when they start, and you keep presence with the 95% who’ll buy later.
It gets bought on its own. Intent on its own tells you an account is researching your space. Combined with first-party high value actions and buying-committee breadth, it tells you who’s researching, how seriously, and when to step up. Most teams never join those up.
Mikelistke: Intent data is some of the best early signal you can get in B2B. Where it goes wrong is how it’s used. The teams that do well treat a surge as the moment to start showing up for that account, with ads, content and a close eye on their own site, and only pick up the phone once the account’s own behaviour says it’s time.
Signal types and how to use each one
Not all signals are equal. This list is my working view, built from the signal stack post, the HVA guide and the account scoring model. Value means how much the signal is worth when it’s used properly. The last column is where teams usually go wrong.
Fit (firmographic / technographic)
Examples: Industry, size, HQ, tech stack
What it tells you: Whether the account could buy
Value: Medium. It’s necessary but says nothing about timing
How it usually gets misused: Treated as the whole targeting plan, built on old classification codes. See the SIC code problemThird-party / offsite intent
Examples: Publisher co-op topic surges, review-site research
What it tells you: Which accounts are actively researching your space, often weeks before they reach you
Value: High, as an early signal
How it usually gets misused: Fired at SDRs as a sales alert or frozen into a static list, instead of triggering always-on account-level ads and being combined with first-party signalsLicensed audience segments
Examples: Vague pre-built “IT decision maker” segments in a DSP, with no account-level detail
What it tells you: A rough profile match
Value: Low
How it usually gets misused: Bought because they’re easy, even though you can’t break delivery down to named accounts. Segment vs SignalIP-to-company resolution
Examples: Reverse-IP site visitor identification
What it tells you: Which company a visit probably came from
Value: Variable. Single-signal IP can be weak
How it usually gets misused: Used on its own: hybrid work makes it shaky. My Reference Index notes IP agreement rates can be as low as 3% on some enterprise accounts (Reference Index)First-party high value actions (HVAs)
Examples: Pricing, integration and security pages; repeat visits; spec downloads
What it tells you: A specific account is evaluating you
Value: High, especially when several people from one account do them
How it usually gets misused: Never defined, or never resolved to accountsAttention
Examples: Time in view, scroll depth, active engagement with ads and content
What it tells you: Whether anyone actually took in the message
Value: Medium to high as a quality filter
How it usually gets misused: Ignored in favour of clicksBuying-committee breadth
Examples: Multiple roles or offices from one account engaging in the same window
What it tells you: A group process is likely under way
Value: High
How it usually gets misused: Not tracked, because nobody built the account view
The pattern is simple. Third-party intent tells you early which accounts to start showing up for. First-party behaviour and committee breadth tell you when to step up. Used together, they’re far stronger than either on its own.
Segments vs signals
A vague segment is a profile someone else guessed at. A signal is something an account actually did. Too often B2B buys generic DSP segments because they’re easy. If you can’t break delivery down to named accounts or committee members, attribution and the board conversation both fall apart. In Segment vs Signal I also argue that a 50,000-account target list is really dozens of different audiences sharing one CSV, and it should be tiered by deal size.
There’s more on the same theme:
Mike’s take: Whatever data you use, ask one question of your media: can you see which accounts, and which people in them, actually saw and engaged with the ads? If you can, the signal is doing its job. If you can’t, you’ve bought a list and you’re guessing.
The signals that actually predict revenue: HVAs and scoring
A high value action (HVA) is a specific behaviour that has historically come before pipeline, such as a deep read of pricing, integration or security pages, a spec-sheet download, or the same account returning three times in a fortnight. In our experience across B2B advertisers, around 95% of accounts seriously evaluating a vendor won’t fill in a form before the decision is materially formed. That’s why form fills are the wrong thing to optimise for (HVA guide).
HVAs work with third-party intent, not against it. Intent shows the research an account is doing across the market, often before it reaches you. HVAs show what it does once it gets there. My account scoring model blends three things:
Fit signals
Weight in the model: Up to 50 points
What goes in: Firmographics such as industry, size and HQ location, plus technographic fitBehavioural signals
Weight in the model: Up to 30 points
What goes in: Website HVAs such as pricing views, case study downloads and contact requestsMultipliers
Weight in the model: Up to 20 points
What goes in: Recency (activity in the last 7–14 days counts more than anything 30+ days old), committee breadth and similar
In that piece I describe seeing a “meh” fit account go from first touch to deal in under 30 days, because four buying committee members hit pricing pages, integration docs and high-intent content in a single week. That’s why behaviour and breadth sit in the model alongside fit and intent.
Go deeper:
Attention as a Buying Signal. In one FunnelFuel test, a cohort with a 0.08% CTR showed 41% higher scroll completion on high-value product pages than the “high CTR” cohort.
Signals you can’t see: the dark funnel
A lot of B2B research happens where no pixel will ever reach: private Slack and WhatsApp groups, conversations over coffee, research behind a rejected cookie banner, and now AI chat windows. In The Xerox Strategy I argue that this is permanent. The answer is to make sure your brand is already in the buyer’s head before the private conversation starts, rather than trying to light up every corner.
Some of it is recoverable, though:
The B2B Dark Funnel Is a Data Problem, Not a Buyer Problem. Buyers still visit vendor websites before they buy. They’re just harder to identify since they left the corporate office.
The Dark Signal Layer: Capturing What DSPs Miss covers the micro-signals consumer DSPs never see, like a VP of IT abandoning a security assessment form and coming back three days later to finish it.
The Account Reporting Layer Brand Advertisers Keep Throwing Away shows how brand campaigns already produce account-level intelligence that most advertisers ignore.
Mike’s take: You’ll never see all of it, and you don’t need to. Intent and your own site data show you which accounts are moving. Always-on advertising makes sure the rest already know you when they start.
Decision guide: what to do with a signal
Third-party intent surge
What it probably means: The account is researching your space, often early
What to do: Move it into always-on account-level ads now and watch for first-party HVAs. Don’t cold-call off the surge aloneIntent surge and first-party HVAs from one person
What it probably means: One researcher is interested
What to do: Personalise ads and content. Alert the account owner, but use a soft touchHVAs from several roles or offices in the same fortnight
What it probably means: A buying group is probably forming
What to do: Prioritise for sales this week, tailor messaging to each role and raise media pressureHigh fit, no signal at all
What it probably means: Not in market, or researching invisibly
What to do: Keep brand presence and frequency sensible. Don’t drop themSignal from a low-fit account
What it probably means: Curiosity, a student, a competitor or a poor match
What to do: Monitor, but don’t spend sales time
Checklist: getting value from intent data
Define a short list of HVAs from your own won deals, so you can combine them with your intent data from day one.
Resolve signals to accounts, not just to individuals or cookies.
Set rules for each signal combination (see the decision guide above) and agree them with sales.
Name an owner and a response time for “act now” accounts.
Weight signals by what they tell you and combine them. Don’t add them up as if they were all equal.
Review scoring against closed-won and closed-lost deals every quarter.
Keep brand pressure on the high-fit accounts that aren’t showing signals yet.
FAQ
What is B2B intent data?
B2B intent data is evidence that a company is researching a topic related to what you sell. It’s collected from publisher networks, review sites or your own website. It’s a powerful early signal, and it works best combined with your own first-party data.
What is the difference between first-party and third-party intent data?
First-party intent comes from your own website and campaigns, so it shows interest in you. Third-party (offsite) intent comes from other sites, such as a publisher co-op, so it shows which accounts are researching your space, often before they ever visit you. The two are strongest together.
What is offsite intent data?
Offsite intent data is research activity captured away from your own website, for example across a network of B2B publishers. It shows you research you’d otherwise never see, which makes it one of the earliest signals you can get. Pair it with first-party signals to judge how far along an account is.
Why don’t some teams get pipeline from intent data?
Usually because nobody owns the step after the signal. Surges sit in a dashboard or a static list, sales calls them too early or not at all, and there’s no always-on advertising to the accounts that surge. Clear rules, an owner and first-party confirmation fix most of it.
How do you get the most from intent data?
Combine it with your own HVAs and committee breadth, put surging accounts into always-on account-level ads straight away, and act quickly on accounts where several people are engaging.
What are buying signals in B2B?
Buying signals are behaviours that change the odds an account will buy. They include pricing-page visits, repeat visits, several roles from one company engaging, and high attention on your ads and content.
What is a high value action (HVA)?
An HVA is a specific behaviour that has historically come before pipeline, such as a deep read of pricing or integration pages. HVAs are more predictive than clicks or form fills because most serious buyers never fill in a form early.
How do you score accounts using intent data?
Combine fit, behaviour and multipliers. In my model, fit is worth up to 50 points, behavioural HVAs up to 30, and recency and breadth multipliers up to 20. Then review the weights against closed deals every quarter.
Is IP-based company identification reliable?
On its own, often not. Hybrid working has scattered buyers across home networks, and single-signal IP matching can be weak. Use it as one input alongside other signals.
What is the 95-5 rule?
It’s the LinkedIn B2B Institute’s finding, based on Ehrenberg-Bass research, that only around 5% of a category’s buyers are in market at any time. The other 95% will buy later, which is why always-on presence matters alongside intent targeting.
Subscribe
I’m Mike Harty. I’ve worked in adtech since 2011: I built PowerLinks, and I’m now Co-Founder and CPO of FunnelFuel, a B2B programmatic and analytics business. The B2B Stack is where I write up what I see on live campaigns, without the vendor spin.