Intent data

You're dealing with intent data as soon as interest is inferred from online behaviour. It comes from two main sources, and they are not equal.

First-party intent data comes from your own channels: who visited your pricing page, who downloaded your guide, who reacts to your posts. You know exactly where it comes from and it is precise. Its limit: it only covers people who already know you.

Third-party intent data is collected by vendors across a large number of sites, such as trade media, review sites and comparison sites. Companies like Bombora or G2 sell this kind of data. It surfaces accounts that have never come across you, at the cost of lower precision.

In both cases the information is usually tied to a company, rarely to an identified person. You learn that an organisation reads a lot about a theme. You don't learn who is reading, or why.

Its best use is prioritisation. On a list of accounts that already match your ICP, move up the ones showing a spike of interest in a topic related to your offer. Intent data doesn't replace targeting. It orders it.

It also helps you pick an angle. If an account reads mostly about data security, your first message can open on that topic. Never write that you detected their interest: you might be wrong, and even when you're right it feels intrusive.

Because the data stops at the company, you still have to find the right person. That's where LinkedIn takes over: you look inside the account for the decision-maker or the user concerned, and that's who you write to.

Start with your own data before buying any. A known account that comes back three times to your pricing page tells you more than a score calculated somewhere else.

The simplest test splits your account list into two comparable halves. In one, you prioritise using intent data. In the other, you ignore it. Same message, same pace, same period. Then compare replies, meetings and signed deals. If the gap is small, the price of the data isn't justified.

Measure precision as well: of the flagged accounts, how many were actually looking for a solution when you spoke to them? That question gets answered during discovery calls, and the answer is worth more than any score.

For an order of magnitude on a related signal, the MimikFlow 2026 Observatory measured the effect of company news (job change, funding round, open role): 40.9% replies against 37.1% without news, and 6.8% meetings against 5.8%. Direct gestures toward the prospect weighed much more: 74.6% replies after a reaction to a post, against 35.6% for strangers found through search.

It is probabilistic. A reading spike on your topic might come from an intern writing a memo, a competitor doing research or a consultant working for someone else. Treat it as a clue, never as a certainty.

It stays at company level. In a 2,000-person organisation, knowing that somebody reads about a theme doesn't tell you who to write to.

Coverage varies. The site networks feeding this data don't cover every country, language or company size equally. Check coverage in your market before you sign.

Finally, ask where it comes from. In Europe, ask the vendor how the data is collected and on what legal basis, and document how you use it.

MimikFlow works with signals you can verify. On one side, direct gestures toward you: a profile visit, a reaction to a post, an invitation received. On the other, public and dated facts: hiring, a new role, a funding round, an expansion. Every signal stays tied to a real event, never to a guess from the model.

Those signals are used to order prospects who already match your target. When the fact is public, the first message can start from it instead of a one-size-fits-all opener.