Why 2nd Party Intent Data Needs an Inference Layer

A case for 2nd party intent data and why it's underlooked.

The term intent data is often used in the context of 1st party like website visits, pricing page clicks or 3rd party black box intent data like Bombora.

But 2nd party intent data has a bad rep of being a commodity because it's often associated to - A funding round. A hire. A job change.

Problem is - it feels un-nuanced and mostly noise.

It lacks relevance to your product and is weak.

And stacking weak independent signals together doesn't make it better.

Funding + hire + job posting isn't a "story." That's just scoring with extra steps.

Real signal depth looks like a story.

A job posting saying "we're building agent evals and monitoring for 99% accuracy in production" and someone just got hired to do exactly that.

Full context. A very specific intent to do things, the pain point, the maturity stage and the person driving it.

2nd party intent needs an inference layer to work.

Raw signals need to be compressed against what your product does to produce a story.

This is what LLMs were built for. And 2nd party data is context-rich and time aware.

Which other signals don't they are individual "hits" website visit+ad click or bombora keyword searched for "AI" - has no context to it.

While building OpenFunnel we naturally started with 2nd party data and trying to make sense of it - as it was in plain sight that LLMs can read and infer.

Now imagine getting pinged daily on "stories" from your entire TAM. Just like news.

When it works you don't get a lead score. You get a relevant story.