Enterprise / PE-backedFintech & Financial Services
Two scoring models were running at once, and good-fit leads were the casualty
Two models became one
Two models became one
Two lead-scoring models were running at the same time, a general model and a campaign-specific one, with no single answer to what a good lead was. The fit component could subtract points, which meant genuinely good-fit, highly engaged leads scored below the routing threshold and never reached a rep's queue at all.
Consolidated the two competing models into a single system with two separately visible scores, engagement and ICP fit, so a lead's behaviour and its fit stay legible on their own instead of collapsing into one opaque number nobody trusts.
Shipped an interim change so the fit component can only add points, never subtract, releasing good-fit, high-engagement leads into the rep queue while the full model was still being built, value in week one rather than at go-live.
Added the platform's native predictive scoring into the model design rather than running it as a competing signal on the side.
Documented the scoring model on a single live enablement site so sales and marketing argue from the same definition instead of two different decks.
Built a 90/180-day reactivation model for leads marked not-ready, with a short reactivation sequence, so disqualified leads get systematically re-touched instead of ageing out of the database.
Two models became one
The general and campaign-specific scoring flows were collapsed into a single dual-score engagement plus ICP-fit model, documented on one live enablement site as the source of truth.
