What Is Deal Scoring?
Deal scoring is a probability that a specific opportunity will close, and close on time, built from the signals in your CRM instead of a rep's gut feel about a close date. A rep-set close date is a single guess, entered by the person with the most incentive to sound optimistic. A deal score reads the actual behavior of the opportunity: how the buyer is engaging, and how the deal stacks up against every similar deal your company has closed before.The fastest accuracy gain in most pipelines comes from one change. Stop treating the close-date field as a forecast input, and start scoring the deal underneath it.
Why Do Rep-Set Close Dates Fail as a Forecast?
Rep-set close dates fail because the date carries no evidence, and the value behind it rarely survives the quarter. In our data, of the pipeline that carries an in-quarter close date on the first day of the quarter, about 20% actually closes inside that quarter. The other 80% of that value does not land when the dates promise it will. A forecast built on those dates inherits every optimistic guess in the pipeline.The strongest single sign that a deal is slipping is one reps create themselves: a changed close date. When a rep pushes a date from one quarter into the next, that deal becomes less likely to close at all, even while it still sits in Commit. The date is not a plan. It is a symptom.
What Signals Should a Deal Score Use?
A deal score should combine two signal families: engagement signals that show whether the buyer is moving, and MEDDIC signals that show whether the deal is qualified. Engagement proves the deal is alive. MEDDIC proves it is real. You need both, because a deal can be busy and unqualified, or well qualified and quietly stalling.| Signal family | What it measures | Example inputs |
|---|---|---|
| Engagement | Whether the buyer is moving the deal forward | Recent stage change, buyer email replies, meeting cadence, days since last activity |
| Metrics | Whether the economic case is quantified | Documented ROI, cost of inaction, the metric the buyer owns |
| Economic Buyer | Whether the budget holder is involved | Direct contact with the person who controls spend, not only a champion |
| Decision Criteria | Whether you know how they will choose | Written requirements, ranked priorities, a scorecard |
| Decision Process | Whether you know the steps to signature | Procurement, legal, and security review mapped to an approval chain |
| Identify Pain | Whether the pain is urgent and owned | A named business pain tied to a deadline |
| Champion | Whether someone sells for you internally | An advocate with access and influence |
How Do Engagement Signals Predict Slippage?
The earliest warning that a deal will slip is not a bad update, it is the absence of any update at all. A deal with no change in stage, no change in amount, no change in close date, and no new notes is more dangerous than a deal with a problem you can name. Buyer silence reads the same way. When a buyer stops returning email and stops taking calls, the deal has already turned, whatever the stage says.We count meaningful activity as a specific short list: a change in stage, close date, or amount. Everything else is noise. Under a 12-month rule, 10% or more of a typical pipeline sits untouched for a full year and still shows up in coverage math as if it were live. A deal score marks that pipeline dead so the forecast stops counting it.
Timing matters as much as movement. At ORM each opportunity is grouped by a machine learning model, and every group gets its own predicted close curve. Those curves run from 1 to 80 weeks, with most of the closing expectation landing before week 12 and very few groups expecting anything past 52 weeks. A deal that has blown past the close curve for its group is off track, no matter what date the rep typed in.
How Do You Turn Signals Into One Score?
You turn signals into a score by weighting each input by how well it predicted closes in your own history, then rolling them into a single 0-to-100 probability per deal. The weights are not opinions. They come from what actually closed. Here is an illustrative pair of deals, both marked Commit by the rep, scored on the same inputs.| Input | Deal A | Deal B |
|---|---|---|
| Days since last meaningful activity | 4 | 47 |
| Close-date changes this quarter | 0 | 2 |
| Economic buyer engaged | Yes | No |
| Decision process mapped | Yes | Partial |
| Pace vs group close curve | On track | 3 weeks past |
| Deal score | 84 | 29 |
How Is Deal Scoring Different From Pipeline Coverage?
Pipeline coverage tells you how much pipeline exists. Deal scoring tells you how much of it will convert. The standard coverage rule of 3x to 5x, and most teams sit near 3.5x, treats every dollar of pipeline as interchangeable. It is not. Coverage says nothing about whether the pipeline is concentrated in the wrong stage, owned by the wrong reps, aged past the point of life, or priced above what deals actually close for. A pipeline can carry an $80,000 average deal size while closed-won deals average $40,000, and the coverage ratio will still look healthy.Deal scoring is what makes coverage honest. It reweights the pipeline by probability, so the forecast reflects the deals that will close and the value they will close at. That is the payoff of moving off the close date. You stop forecasting the pipeline you can see and start forecasting the pipeline that is going to happen.
At ORM we build the models that do this. Each opportunity is grouped by machine learning, scored against the close curve for its group, and rolled into a forward forecast. The models train on 4 to 6 weeks of your historical sales performance, and the forecast they feed targets 95% accuracy and holds from day 1 to day 90 of the quarter rather than sharpening only in the final week. The scores surface in Radar, our in-app AI, and stay queryable from whatever LLM you connect. Move off the close date, and the quarter stops being a surprise.
Frequently Asked Questions
What is deal scoring?
Deal scoring is a probability that a specific opportunity will close, and close on time, calculated from CRM signals instead of a rep's close-date guess. It reads engagement, such as recent activity and buyer responsiveness, alongside qualification signals like MEDDIC, then compares the deal to how similar deals closed before. The output is a single score per deal that a forecast can trust more than a date field.
Is deal scoring better than a rep-set close date?
Yes. A close date is a single optimistic guess from the person most motivated to sound positive, and in ORM's data the strongest sign a deal is slipping is a rep changing that date. Once a deal slips from one quarter into the next it becomes less likely to close at all, even while it still sits in Commit. A deal score replaces the guess with evidence from the deal's own behavior.
What signals go into a deal score?
Two families. Engagement signals show whether the buyer is moving the deal: recent stage, amount, or close-date changes, meeting cadence, and days since the last real activity. MEDDIC signals show whether the deal is qualified: metrics, economic buyer, decision criteria, decision process, identified pain, and a champion. Engagement proves the deal is alive, MEDDIC proves it is real, and a good score needs both.
How does deal scoring use MEDDIC?
MEDDIC gives the score its qualification backbone. Each element, from a quantified metric to an engaged economic buyer to a mapped decision process, becomes an input the model weights by how well it predicted closes in your history. A deal missing its economic buyer or its decision process scores lower even when the rep is confident, because those gaps are where confident deals stall.
What is the earliest sign a deal will slip?
The absence of a signal. In ORM's data the earliest warning is not a bad update, it is no update at all: no change in stage, amount, or close date, and no new notes. Buyer silence works the same way. When a buyer stops returning email and stops taking calls, the deal has already turned, whatever the stage says. A deal untouched for a year still shows up in coverage math, and a deal score is what marks it dead.
See how ORM turns these insights into action
ORM builds custom revenue forecast models for B2B SaaS companies. Not dashboards. Prescriptive analytics that tell you what to do next.
Schedule a Demo