Intent from your own ads vs third-party intent data
First-party intent from your LinkedIn and ChatGPT ads, website and AI answers versus bought third-party intent data. What each one actually measures, where each one misleads, and how to combine them in ABM.
First-party intent is what companies do with your own ads, pages and AI answers: which themes they engage with, which pages they read, which buyer questions brought them to you. Third-party intent is research activity observed elsewhere on the web and sold to you, usually as topics a company is "surging" on. First-party intent is narrower but specific to you and easy to explain to sales. Third-party intent is broader but noisier and harder to act on. Most B2B teams should build on first-party intent and treat third-party data as a prospecting input, not a trigger.
What third-party intent data measures
Third-party intent providers collect content consumption across a network of B2B websites and map it to companies, usually by IP address. Bombora, one of the best-known providers, describes a "Data Co-op" of thousands of publishers, brand sites and data providers, with 25,300+ intent topics and Company Surge data reflecting "14B+ monthly events and 5M company domains" (Bombora), figures as stated on its site at the time of writing. In general terms, a company is flagged as "surging" on a topic when its research on that topic rises above its usual level.
That is genuinely useful at the top of the funnel: it can surface companies you'd never have put on a list, researching a category you sell into.
It has three structural limits:
- It's topic-level, not vendor-level. A surge on "endpoint security" means someone at the company read about endpoint security. It doesn't mean they're evaluating you, or buying this quarter.
- It's company-level by IP. The same reverse-IP limits apply as with any visitor identification: remote workers on home broadband and people on mobile data aren't attributed. Large companies with many offices generate surges constantly; small companies rarely do.
- You can't see the evidence. You get a topic and a score. You usually can't tell sales which pages or what was read, which makes it hard to write a relevant first message.
What first-party intent measures
First-party intent comes from your own channels, where you control the content and know exactly what the signal means.
| Source | What it tells you | Strength |
|---|---|---|
| LinkedIn ad engagement by company | Which accounts engaged with which campaign theme | You know the theme because you wrote the ad |
| Website pages read | What they researched on your site | Specific to your product and pricing |
| High-intent pages | Pricing, demo, contact | The strongest signal you'll get without a form |
| ChatGPT Ads clicks | Which buyer-question ad group they clicked | Shows the kind of problem they were asking about |
| Visits from AI answers | ChatGPT, Perplexity or Gemini sent them to you | They were researching your category in an AI assistant |
| CRM deals | Whether there's already an active process | Prevents false "hot" alerts on accounts already in a deal |
The key advantage is interpretability. "Engaged with our [Security] campaign three times and read the SOC 2 page" is a sentence an SDR can use. "Surging on Information Security, score 72" isn't.
LinkedIn engagement is intent if you design for it
LinkedIn reports impressions, clicks and engagements per company. On their own, impressions are your spend, not their intent. Engagement with a specific theme is different: if you run separate campaigns for separate themes, the theme an account engages with is a first-party intent signal.
That only works if campaign structure reflects themes. One campaign with mixed creative tells you an account engaged, not what it cared about. Three campaigns named "ABM | [Security]", "ABM | [Integrations]" and "ABM | [Pricing]" tell you what. In IntentBid, a theme in square brackets, a hashtag or "Intent: Theme" in the campaign name becomes an intent tag on every account that engages, with no setup. Page rules add tags from what accounts read on your site.
The limits of first-party intent
Be honest about the gaps:
- It only sees accounts you already reach. If a company isn't on your LinkedIn list and doesn't visit your site, it has no first-party intent. That's where third-party data earns its keep.
- LinkedIn's company data is approximate and partial. LinkedIn states professional demographic metrics are approximate, values with fewer than 3 events are dropped, and only the top 100 values per creative per day are returned (LinkedIn Marketing API). Light engagement at a small company can be missing.
- Website identification is partial. Visits from home broadband, mobile data and VPNs can't be tied to a company by anyone. Typically 15 to 30% of business visits can be identified; see Why some visits have no company.
- It's biased by your own spend. Accounts you spend more on will show more engagement. Normalise by impressions (engagement rate) before comparing accounts.
Where each one misleads
| Situation | Third-party intent says | First-party intent says | What's usually true |
|---|---|---|---|
| Big enterprise, many offices | Surging on 12 topics | Low engagement rate | Lots of reading, little interest in you |
| Small company, remote team | No surge | Read pricing twice via ChatGPT | Real, early evaluation |
| Customer | Surging on your category | Many visits | Existing user reading docs, not a new deal |
| Competitor | Surging on your category | Visits to pricing | Competitive research |
| Analyst or agency | Surging | Engaged with every campaign | Not a buyer |
The last three rows are why exclusions matter for both kinds of intent. Remove customers, competitors and your own company before you score anything.
How to combine them
A practical split that works for most mid-market B2B teams:
- Use third-party intent to build and refresh the list, not to trigger outreach. Companies in your ICP surging on your category are good candidates to add to the target account list for the next quarter.
- Use first-party intent to move accounts through stages. Engagement with themed campaigns, high-intent pages, ChatGPT ad clicks and AI-answer visits decide whether an account is Interested, Considering or Selecting.
- Trigger sales on first-party signals only. A rep calling because of a topic surge has nothing specific to say. A rep calling because the account read the integrations page after engaging with the integrations campaign does.
- Route themes into messaging. If accounts tagged "Security" convert better, make the next LinkedIn campaign for Interested accounts about security, and write the outreach about it too.
If you don't buy third-party intent at all, you're not missing the core of ABM. You're missing a list-building input, and you can partly replace it with lookalikes of your best customers and with the buyer questions you track in AI answers.
Measure intent by outcome, not volume
Whichever source you use, test it the same way. At the end of a quarter, take each intent tag or signal and ask: of the accounts that showed it, how many created an opportunity? Compare that with accounts that didn't.
The Which intents move accounts forward table in IntentBid's Account engagement report does this per intent tag: accounts, accounts with a deal, and pipeline. Themes that never correlate with deals are either the wrong message or the wrong signal.
How IntentBid helps
IntentBid builds first-party intent from your own LinkedIn and ChatGPT Ads, your website and AI answers: intent tags from campaign names and page rules, stages from your own conditions, and both sent to HubSpot daily. It doesn't sell third-party intent data. See Set up ABM stages and intent tags and the pricing.
FAQ
Is first-party intent the same as website visitor identification?
Visitor identification is one source of first-party intent. Ad engagement by company, ChatGPT ad clicks, AI-answer visits and CRM deals are others. Visitor identification alone misses everyone who engages with your ads but hasn't visited yet.
Can I get first-party intent from ChatGPT Ads?
Partly. ChatGPT Ads don't report by company, and OpenAI doesn't share the conversation with advertisers. But if each ad group covers one buyer-question cluster and its landing page URL carries UTMs per ad group, the click tells you which cluster it came from. If the visitor's company can be identified on your site, you know which company clicked on which kind of question.
Is third-party intent data GDPR-compliant?
It depends on the provider and how it collects and processes data. The EU Court of Justice held in the Breyer case that dynamic IP addresses can be personal data (Bird & Bird). Ask any provider for its legal basis and data processing terms before you buy.
How many intent themes should we run?
Three to five at a time. More than that splits the LinkedIn budget so thinly that each theme gets too few impressions per account to produce a readable signal.
Do intent tags replace account scores?
No. Tags say what an account is interested in; scores and stages say how far along it is. Use both.