Intent data is sold harder than almost anything else in B2B marketing. It is also understood worse.
The pitch is attractive: here is a list of companies researching what you sell, go call them. If you run an account-based marketing programme, that list looks like the missing piece — the thing telling you which named accounts to work this month.
Sometimes it is. Often the data cannot carry the weight put on it. This page is about telling those cases apart before you sign.
What intent data actually is
Intent data is observed behaviour, converted into a guess about buying stage.
A vendor's pricing page gets four visits from one company in a week. A review site logs a comparison between two products. None of that is a stated intention. All of it correlates with being in-market.
The useful signals fall into four groups:
- Content consumption — topic pages, guides, whitepapers, webinars.
- Comparison behaviour — review sites, alternatives pages, competitor pricing.
- Commercial behaviour — your own pricing page, demo forms, documentation.
- Technographic change — a tool added or dropped, a job posting that names a system.
Everything else in the category is packaging. The question is never "is this intent data" — it is "who observed this, where, and at what resolution".
Account-level, not person-level.
The three sources, side by side
Every intent product is built from one of three supply types, or a blend. These differences matter more than any feature list.
| Source | What it observes | Resolution | Freshness | What it proves | What it does not |
|---|---|---|---|---|---|
| First-party your own properties |
Pages, forms, emails, product usage, call records | Known people where they identify themselves; company by IP otherwise | Real time | Interest in you, specifically | Anything happening off your site |
| Second-party publisher and review co-ops |
Reading and comparison on a defined set of B2B sites | Mostly company-level; person-level only where members registered | Daily to weekly | Category research is rising at an account | Which person, which budget, which timeline |
| Third-party brokers, bidstream, content syndication |
Ad auction requests, aggregated feeds, scraped or licensed sets | Device or IP, inferred to a company | Varies; often opaque | That some activity was seen somewhere | Provenance, consent, or that the company is yours to call |
Two named examples make the second-party row concrete. Bombora says its data is "directly collected, with consent" from a co-operative of "5,000+ sites", resolves signals to businesses, and argues against bidstream as a source (Bombora, Our Data). G2's documentation states that for Buyer Intent "all data is collected at the organization level", and lists the page types it counts — your profile, pricing, alternatives and competitor comparison pages (G2 documentation, accessed 15 Oct 2026).
Informa TechTarget is the exception on resolution. Its Priority Engine material says its intent "comes from registered, opt-in, active people" — person-level, because they logged in to get the content (Informa TechTarget, 2023).
The bidstream question, handled honestly
Bidstream intent is built from the data inside programmatic ad auction requests.
When a page loads, an auction request goes out carrying a device identifier, an IP address and a page URL. Hundreds of bidders see it. A bidder that never wins still saw the request. Some firms kept and resold that view.
That practice is contested on two fronts.
- Platform rules. Google's Ad Manager data collection control lets publishers decide whether bidders may use bid request data to build profiles and remarketing lists. With it off, "bidders from the Authorized Buyers program are not allowed to build user lists and profiles based on data from your traffic" (Google Ad Manager Help, accessed 15 Oct 2026).
- Regulators. In December 2024 the US FTC acted against the data broker Mobilewalla, alleging it "collected data from real-time bidding exchanges and third-party aggregators" and "unfairly collected and retained the information in the bid request, even when it didn't have a winning bid" (FTC, 3 Dec 2024).
Read that carefully. It does not mean every bidstream-derived product is unlawful. It means the supply is narrowing and the legal position is unsettled, so a product built mainly on it has an input that could shrink without notice. If a provider cannot name where a signal was observed, assume bidstream until proven otherwise.
Account-level or person-level: the distinction the pitch blurs
This is the most important thing to understand, and the one most often left vague in a demo.
Account-level means something at this company looks interested. Not who. Not how senior. Not whether they hold budget. 6sense frames its own intent this way, saying it connects searches "back to the accounts that are making them" (6sense, accessed 15 Oct 2026).
Person-level means a named individual. It exists in two honest forms: people who registered on a publisher's property, and visitors your own site already knows. Everything else is inference.
Geography matters enormously for Indian B2B. RB2B, one of the better-known person-level visitor identification tools, states on its own site that "Person-Level ID is US-only technology with a US-only database" and that "it will not identify individuals located outside the US", while company-level identification is global (RB2B, accessed 15 Oct 2026).
So if your buyers sit in Mumbai, Bengaluru or Gurugram, person-level website identification is mostly unavailable to you. Plan for account-level, and work the account with research rather than a handed-over name.
The first-party signals most Indian B2B companies already ignore
Here is the uncomfortable part. Most companies asking us about buying intent data are not using the intent data they already own. These signals are free, unambiguous and about your product rather than a category:
| Signal you already have | Why it beats bought data | Where to find it |
|---|---|---|
| Pricing page visits | Nobody reads pricing casually | GA4 page reports, server logs |
| Repeat visits from one company in a short window | Several people are looking, which means a committee | Reverse-IP tooling, CRM web activity |
| Documentation or integration pages read | Evaluation, not awareness | GA4, docs analytics |
| Form started and abandoned | Intent plus friction — the cheapest fix you own | Form analytics, event tracking |
| Demo or contact page bounces | They came to buy and something stopped them | Exit-rate reports, session recording |
| Old opportunity re-engaging | A real buying cycle restarting | CRM plus email engagement |
Getting this layer working is a first-party data job, not a procurement job. It needs clean event tracking, those events landing in the CRM, and a marketing automation rule that puts a pricing-page visit in front of a salesperson the same day.
Bought intent tells you a category is warming. Your own data tells you someone is choosing.
DPDP and intent data in India: a design question
This is a design consideration, not legal advice — take the legal reading from your counsel.
India's Digital Personal Data Protection Rules, 2025 were notified on 14 November 2025 with an eighteen-month phased compliance period, and they require a separate, clear consent notice stating the specific purpose for which personal data is collected (PIB, 17 Nov 2025).
Three things follow for intent work:
- Company-level signals are structurally easier. "Activity rose at Acme Ltd" is not a person. Person-level identification of Indian individuals is where consent questions begin.
- Provenance becomes a procurement requirement. If a vendor hands you named Indian contacts, you need to know on what basis.
- Your own consent notice must match what you do. If you score and route behaviour, say so in the notice.
Our walkthrough of the obligations marketers actually touch sits in the DPDP Act for marketers.
The questions that expose a weak provider
Demos are built to avoid these. Ask them anyway.
- Where was this observed? Name the properties. "A proprietary network" is not an answer.
- Is any of it bidstream-derived? A straight no, or a straight yes with a percentage, both beat a deflection.
- Account-level or person-level — for Indian accounts specifically? Many tools degrade sharply outside the US.
- What is the baseline? A surge only means something against that company's own normal. Without a baseline you are reading traffic, not intent.
- How many of my target accounts can you report on? Run the overlap test below.
- How old is a signal when I see it? Weekly refresh on a four-week sales cycle is useless.
- How is consent obtained, by whom? Co-ops put this on the publisher. Get it in writing.
- What happens at renewal if supply shrinks? Ask for the clause, not the reassurance.
What we tell a B2B client who asks about buying intent data
This is DH's own view, from running organic and paid programmes for B2B accounts in India. We have no client results to quote here — what follows is the sequence we recommend.
First, we ask what would change on Monday. If nobody has time to act on thirty warming accounts, the data is not the bottleneck. Capacity is.
Second, we instrument first-party signals before anything is bought. Pricing page, documentation, form abandonment, repeat company visits. This usually takes a few weeks and finds demand that was already there.
Third, we check coverage, not features. Indian mid-market accounts are thin in most global intent datasets, so we run the overlap test above. If coverage is poor, no amount of scoring sophistication saves it.
Fourth, we decide what the signal will trigger. Usually a nurture sequence and a tighter ad audience, not a cold call — an account-level signal does not justify ringing a named person who never raised a hand.
Fifth, we buy small and time-boxed. One quarter, one segment, one question: did accounts flagged as surging convert better than matched accounts that were not? That comparison is the only honest justification for a renewal.
Most of the value we have seen sits in steps two and four. Those cost nothing but discipline.
Where intent data earns its place
Used well, intent data does one job: it changes the order of your work.
It will not tell you who to call, and it cannot substitute for a B2B marketing strategy that explains why anyone should switch. What it can do is tell you which thirty of your three hundred named accounts deserve this month's attention. That is genuinely valuable if you have the capacity to act on it.
Treat it as a prioritisation input with known limits. Build your own signals first, because they are better, cheaper and yours. Then buy narrowly, test against a control, and renew only on evidence.
The next layer is a B2B lead generation engine that can act on the signal, and an ABM plan that decides which accounts matter in the first place.
Frequently asked questions
What is intent data in simple terms?
It is observed research behaviour turned into a guess about buying stage. Someone at a company reads about a category, compares products or visits a pricing page. A provider collects that activity and reports that the company may be in-market. It is a signal of interest, not a stated intention.
What is the difference between first-party and third-party intent data?
First-party intent is behaviour on your own properties — site, emails, product, CRM. You own it, it is real time, and it is about you specifically. Third-party intent is bought from brokers who observed activity elsewhere, usually resolved to a company by IP or device, with provenance you cannot verify.
Is intent data accurate?
We will not quote an accuracy figure, because no provider publishes one you can independently verify. Accuracy depends on three checkable things: whether a baseline exists for each company, how well the provider covers your actual target accounts, and how fresh a signal is when you see it.
Does intent data tell me which person to contact?
Usually no. Most products resolve to a company, not an individual. Person-level data exists where someone registered on a publisher's property or identified themselves on your own site. For Indian accounts it is mostly unavailable, so plan to work the account rather than a handed-over name.
Do Indian B2B companies need intent data providers at all?
Many do not, at least not first. Coverage of Indian mid-market accounts is thin in most global datasets, and those companies usually have unused first-party signals — pricing visits, repeat visits from one company, documentation reads, abandoned forms. Instrument those, then buy narrowly if a gap remains.
Signals are only worth what you do with them
We build the tracking, audiences and campaigns that turn warming accounts into pipeline.
Make Digital Hangover a preferred source
One tap tells Google to show more of our SEO and marketing coverage in your Top Stories.