A scoring model that sales ignores is worse than no model, because it still generates confident-looking MQLs that reps have privately written off. Here are three models that reliably get quietly abandoned — and the specific reason each lost the room.
Model one: the everything-counts score
This one awarded points for every conceivable interaction. Opened an email, five points. Visited the blog, five points. Attended a webinar, five points. The logic was “engagement is good”, so everything scored.
Why sales ignored it: a lead could hit the MQL threshold purely by being on a mailing list and idly clicking, with zero buying intent. Reps burned through a few of these, learned the score meant nothing, and started ignoring it wholesale. The model measured attention, not intent, and the two aren’t the same.
The fix: strip it to the signals that actually preceded won deals, weight them hard, and bin the long tail of five-pointers.
Model two: the fit-only score
This one scored purely on firmographics — company size, industry, title. A perfect-fit company scored high whether or not anyone there had done a single thing.
Why sales ignored it: it flagged dream accounts who’d shown no interest, so reps chased ghosts. High score, cold lead, wasted call. Fit without engagement is a target list, not a scored lead, and pretending otherwise cost reps their time.
The fix: score fit and engagement separately, and require both to clear a bar before a lead is sales-ready.
Model three: the never-validated score
Technically fine — reasonable signals, sensible weights. But it was built once, at launch, and never checked against outcomes. Eighteen months later the business had changed and the model hadn’t.
Why sales ignored it: high-scoring leads had stopped converting better than low-scoring ones, and reps noticed before anyone in ops did. Once the score demonstrably didn’t predict, its credibility was gone.
The fix: validate quarterly against real conversions. A model that isn’t re-checked is decaying whether you look or not.
The common thread
Every ignored model failed the same test: it didn’t predict revenue, or nobody proved that it did. Reps are ruthless empiricists — they’ll trust a score exactly as far as it’s earned. Build on evidence, separate fit from interest, and keep validating, or watch it quietly become wallpaper.
Related reading
See building HubSpot lead scoring from zero and six signs your lead scoring is being ignored.
Suspect yours is being ignored? Book a health check.