Most scoring models are invented in a marketing offsite and ignored by sales within a month. We build ours backwards from your closed-won deals, then review the model against outcomes every quarter.
In short
Build lead scoring from evidence, not opinion: analyse which attributes and behaviours your closed-won deals actually shared, weight fit and intent separately, keep the model small enough to explain in one sentence, and review it quarterly against outcomes. Pair it with routing rules and SLA timers so a high-scoring lead reaches a named rep within minutes.
Signals
Reps work their own list because the scored leads did not convert last quarter.
Nobody owns the follow-up window and good enquiries go cold over a weekend.
When the assigned rep is on holiday, the lead simply stops.
It was set up two years ago against a product that has since changed.
What you get
We analyse closed-won versus closed-lost to find the attributes that genuinely correlate, then build the model on those.
ExploreRules for territory, segment or account ownership, with round-robin, capacity caps and out-of-office fallbacks.
ExploreTimers on first response, alerts when they breach, and a report that shows who is actually meeting them.
ExploreA quarterly review comparing score bands to real conversion rates, and a mechanism for reps to flag bad leads.
ExploreHow it runs
Look at what closed-won leads had in common and, just as importantly, what the losses had.
Score the last two quarters retrospectively. If the model would not have surfaced the deals you actually won, it changes.
Scoring, assignment, timers and alerts built together, because a score with no routing behind it changes nothing.
Conversion by score band, rep feedback and a recalibration. Models decay; scheduled reviews stop them decaying silently.
Questions
The questions we get asked on almost every call about this work.
Predictive and multiple custom scores need higher tiers, but a single well-built score property with workflows works on Professional. We design to the tier you have.
Few enough that a rep can explain why a lead scored highly. In practice that is usually eight to fifteen weighted criteria, split across fit and intent.
Then we start with a fit-only model built on your ideal customer profile, keep it deliberately simple, and add behavioural weighting once there is data to fit against.
Assignment should be instant and the first-response SLA is typically measured in minutes for inbound demo requests. We build the timers and the reporting to hold that.
Yes — pooled queues, territory teams and account-based ownership are all supported, with capacity caps and fallbacks.
Next step
Give us access to two quarters of closed deals and we will show you what your winning leads actually had in common.