We retired the single intent score. Fit asks whether you should sell to a company at all. Signal asks why now, and its evidence loses half its weight every 60 days. Together they set a tier, and the tier decides your week.
Why one score was retired
Early versions of Caicle produced one number per company. It blended how well the company matched your profile with how much had recently changed there, and it was tidy to sort by. It was also impossible to act on. A perfect-fit company with nothing happening and a poor-fit company with a funding round could land on the same number, and the rep had no way to tell which sentence to write.
The two questions are different in kind. Fit is a judgement about you and the company: is this someone you should be selling to? Timing is a judgement about the calendar: is this the month to do it? Blending them threw away the very distinction a rep needs. So we split them.
Fit: renormalised over the criteria you set
Fit is computed against your ICP, criterion by criterion, and every point is explained: industry matches logistics, 25 of 25; 240 employees is inside the target band, 20 of 20; located in Hamburg, 20 of 20. Only the criteria you filled in are scored, and the result is renormalised over them. A profile with three criteria set produces a score out of the same 100 as one with eight.
That is what makes a profile a yardstick rather than a filter. Leave a criterion blank and it drops out of the score instead of excluding good accounts. Fit is null without a profile; it never pretends to be a judgement it has no basis for.
Signal: dated evidence with a 60-day half-life
Signal is built from dated events with a source: a new managing director in the register, three sales roles posted, a new applicant tracking system, a change of parent in the LEI register, coverage of a funding round. Each signal contributes a weight, and that weight halves every 60 days. A round announced last week counts in full; the same round announced four months ago counts for a quarter.
Signals saturate rather than cap. The tenth signal still counts, a little, so a company with a great deal happening scores above one with a single loud event, but neither runs off to infinity. Every point is attributable to a piece of evidence with a URL, and re-running research refreshes the list without disturbing your notes or pipeline state.
A funding round should not prop up a stale list. Decay is how the pipeline forgets on its own, so that you do not have to remember to clean it.
Tier: two scores, one decision
Fit and signal together set a tier, and the tier is what ranks the Overview. Strike is high fit and strong timing: right company, right month. Engage is high fit with quieter timing: the right company, worth a considered approach rather than a rush. Nurture is a company that could become a fit or is showing movement without matching the profile yet: keep it warm, keep researching. Watch is everything else you have chosen not to dismiss.
Tiers are recalculated as signals decay, so a strike account that goes quiet drifts to engage without anyone touching it, and a nurture account that hires a new commercial lead rises on its own.
What each tier means for a rep's week
Strike accounts get the Strategy Room. Two or three at a time, a plan for your review, approved and delivered this week. Engage accounts get a dossier read and a single well-placed touch that references the fit rather than an event. Nurture accounts get a sentiment check and a note to re-run research in a month. Watch accounts get nothing, on purpose.
That allocation is the whole point of keeping the questions apart. A blended score told you which companies were interesting. Two scores and a tier tell you what to do about each of them on Monday morning, and why the answer will be different next month.