Behavioral Lead Scoring with Decay Automation: A Score That Fades When Attention Does

A scoring model that only ever adds points rewards history, not current interest — a contact who opened five emails and visited the pricing page three months ago can end up ranked above someone who did the exact same thing yesterday, simply because the older contact's total has had longer to accumulate. The score stops answering the question a sales team actually needs answered, which is not "how much has this contact ever done" but "how engaged does this contact look right now."

Deciding how many points a behavior is worth is the visible, easy part of a scoring model. The harder part is deciding how that value should fade — how many days without a repeat behavior before a rule's contribution starts shrinking, and by how much — because getting it wrong in either direction breaks the score's usefulness: decay too aggressively and a genuinely engaged contact looks cold after a brief quiet spell; decay too slowly and the score becomes exactly the stale, ever-rising number that misleads a sales team in the first place. And a scoring model without a cap on how many times a single rule can apply to the same contact lets one repeated, low-value behavior drown out everything else.

A team defines scoring rules against real tracked behaviors — visiting a specific page, opening an email, clicking a link, submitting a form — each with its own point value (positive or negative), its own decay period in days, and its own maximum number of times it can apply to a single contact. When a tracked event happens, the matching rule's points are applied to that contact's running score, respecting the rule's own application cap.

A daily scheduled job then walks every scored contact and reduces the contribution of each rule based on how long it has been since that specific behavior last occurred for that contact, using the rule's own configured decay period — a rule with a short decay window fades quickly if the behavior does not repeat, while a rule with a longer decay window holds its value longer. The team separately configures the score thresholds that map a numeric result to cold, warm or hot, so the labels reflect that specific funnel's own definition of engagement rather than a fixed, one-size-fits-all cutoff. The whole pipeline runs automatically as part of Flow's own contact-event processing — a contact's score updates itself as behaviors happen and fades on its own schedule, with no manual recalculation step for a team to remember.

A contact visits the pricing page, opens two emails and clicks a demo-request link within the same week, pushing their score comfortably into "warm." Sales follows up, but the contact goes quiet for six weeks. Because the pricing-page-visit and email-open rules each carry their own decay period, the daily job gradually reduces their contribution the longer the contact stays inactive — by week six, the score has faded back toward "cold," accurately reflecting that the earlier burst of interest did not continue. When the same contact returns and revisits the pricing page, that fresh event lifts the score again, on top of whatever the decayed baseline still contributed.

A "hot" label starts meaning "engaged recently," not "engaged at some point in the past," letting a sales team prioritize by current interest rather than historical accumulation. Because the thresholds and decay periods are configured, not fixed, a team can tune what "warm" means for its own funnel instead of inheriting someone else's default.

A team that wants a single universal score benchmarked against other companies' funnels will not find that here — thresholds and rule weights are configured per team, with no cross-organization comparison. A team expecting the score to automatically feed a separate CRM pipeline should confirm that connection is actually configured — Flow's scoring pipeline operates on its own contact records, and this page does not claim an automatic hand-off into another Centriu product's pipeline.

A scoring model that only adds points ranks contacts by how long they have existed in the system as much as by how engaged they currently are — an old accumulation can outrank a fresh one. Centriu Flow's decay mechanism keeps the same simple idea of adding points for real behavior, but lets each rule's contribution fade on its own configured schedule, so the number a sales team sees is closer to "how engaged does this look right now" than "how much has ever happened."