Fit tells you who could buy; signals tell you when to pay attention

Most research stacks answer one question: does this company look like my customer? That is a fit filter, and it is necessary but not sufficient. A company can match your ideal customer profile precisely and still have nothing changing that would make a conversation useful this quarter.

The second filter is timing, and it comes from observable change rather than from attributes. A company that just opened a second site, published a new compliance commitment, or reorganised a function has a reason to be re-evaluating things. One that has looked the same for three years usually does not.

So the practical model is two filters in sequence, not one. Fit narrows the universe of companies worth understanding. Signals tell you which of them to prioritise this week.

Paid intent platforms sell anonymised behavioural data at enterprise prices. You do not need that to prioritise a list. Most usable signals are public, and several are visible on the same company website you are already reading during qualification:

• Hiring — open roles in the function you serve imply both budget and a gap
• Leadership change — a new head of a function typically reviews how things are done
• Funding, expansion, or a new site — visible capacity or market change
• Stated priorities — a published roadmap, certification push, or compliance deadline
• Technology and platform changes — a migration or new integration mentioned publicly
• How a company describes itself this year versus last — repositioning is a signal

B2BLead is built around the fit half of that pair, and it is deliberate about it. Under My Business → Products you describe what you sell and who buys it, then Prepare with AI turns that into a scoring brief. Product warm-up ranks real directory companies in your chosen countries against that brief in the background, so ranked Lead Search results already reflect your ICP before you spend a credit.

That gives you an ordered shortlist rather than an alphabetical export. The signal pass is what you do to the top of that list: open the site, read the careers and news pages, and decide whether anything has changed enough to justify attention.

Then run live AI validation on the accounts that survive both filters, reveal contacts only for those, and keep them in Lead CRM with the reason recorded. Credits go to accounts with both fit and a reason — not to an entire industry.

The failure mode of signal-based research is treating it as a platform purchase rather than a habit. A workable loop is small enough to run in one sitting:

• Pull the top ranked accounts from Lead Search for your industry and country
• Spend twenty minutes scanning for the signals above; keep the ones where something changed
• Run live AI scoring on those to sanity-check fit against your current product brief
• Reveal contacts only for keepers, then save to Lead CRM with the signal in the note field
• Next week, start from CRM: which accounts progressed, which went quiet, which signal types actually mattered

Recording the signal that made an account interesting gives you two things. First, any later conversation has context — you can refer to what changed rather than starting from nothing. Second, after a quarter you can see which signal types actually preceded real opportunities for your product, and stop spending time on the ones that did not.

That is the difference between a research habit and research theatre. Teams that abandon signal-based work usually never measured which signals mattered, so it stayed a feeling rather than a method.

It is also the record that lets you explain your own actions later, which matters for its own reasons — see permission-based B2B follow-up for the obligations that attach once you move from researching a company to contacting a person there.

Do not treat a signal as a conversational gimmick. Referencing a company's funding round and then pivoting into an unrelated pitch is worse than no personalisation, because it proves you had the context and ignored it.

Do not stack signals until nothing qualifies either. One good reason is enough. If you need four overlapping signals before an account clears the bar, your fit filter is probably too wide, and the real fix is upstream in your product brief rather than in the signal pass.