Top Platforms for Tracking Deal Health and Risk: What the Score Should Be Made Of
The deal sat at a health score of 82 for six weeks, and when it slipped nobody could say what the 82 had been measuring.
The green score that slipped
Picture the post-mortem. Activity on the deal was high, with a meeting most weeks and a steady stream of email. The score read all of that as health.
What it never registered was that the economic buyer stopped attending after the second meeting, and the champion's replies got shorter each week. Those signals were in the conversations the whole time. The score was built from what was easy to count.
Email volume, meeting count and days since last touch are real signals. They measure motion, and a deal can move a lot while it is losing.
📊 77% of B2B buyers say their latest purchase was very complex or difficult. — Gartner |
Inputs, and what each one can tell you
Before comparing platforms, list what a deal health score could be made of. Each input answers one question about the deal and stays silent on others.
Input | What it tells you | What it misses |
|---|---|---|
Activity volume | The deal is moving | Who is doing the moving |
Days since last touch | Whether the deal went quiet | Whether the right person went quiet |
Stage and close date | Where the rep believes the deal is | Whether the buyer agrees |
Qualification status | Which MEDDPICC letters are covered | Whether coverage rests on evidence |
Stakeholder sentiment and influence | Whether the people who decide are for you | Changes between meetings, if read only once |
Slippage history | Whether committed dates hold | Risk before the first push |
A score built on the first two rows is an activity monitor. The rows further down need someone, or something, to read the conversations.
Where the platforms differ
Forecasting and RevOps platforms such as Salesloft, which Clari is merging into, are strong at rolling deals up into a number and showing how it changed. Conversation intelligence platforms such as Gong hold the calls, which is where the stakeholder and sentiment signals live. Your CRM holds stage, close date and whatever the rep entered.
Each of these can show a health indicator. The difference that matters for risk is how much of the score comes from buyer evidence, and whether you can open the score and see which input moved it.
Questions to put to any deal health score
Can you see the components?
A single number with no breakdown can only be trusted or ignored. Ask the vendor to show one deal's score split into its parts, with the value of each part.
Does it penalize what is missing?
No identified champion and no engaged economic buyer are two of the strongest risk signals in an enterprise deal. A score that only adds points for activity will rate a deal with neither as healthy, as long as someone keeps emailing.
Does it keep evidence apart from opinion?
If the rep's forecast category feeds the score with the same weight as a buyer-confirmed close date, the score repeats the rep's confidence back to you. That helps nobody, the rep included.
Does time on stage change the number?
Most tools can flag a deal that has sat in a stage too long. Fewer let that time change the deal's probability, which is what moves the forecast.
📊 Half of RevOps teams can't name an owner when an AI-generated number turns out to be wrong. — Vasco, 2026 State of AI in RevOps |
What the Spotlight Score is made of
Spotlight.ai calculates the Spotlight Score from evidence captured by Spotlight's Calibrated-Decision AI-Model (CDAM). We didn't train a model for it. CDAM answers narrow questions inside a closed answer space, every answer carries a confidence score and only counts once it clears the bar, and the score itself is calculated in code.
Each component is scored from minus 1 to plus 1, then combined as a weighted adjustment to the deal's base stage probability. The components cover deal progress, flagged gaps and risks, MEDDPICC status, stakeholder sentiment, engagement and influence, the rep's historical win rate, discovery and value depth, deal velocity, how closely the deal matches known winning and losing patterns, recency of meetings and email, and slippage history.
Stakeholder scoring penalizes a deal with no champion or no economic buyer. Time on stage is measured against your own historical average for each stage, and it adjusts the probability, so it flows into the bottom-up forecast. The weights start as heuristics and can move to machine learning over time.
Because the score opens into its parts, a rep can show exactly which input moved and what the buyer said behind it. Gary Olson, CRO at Sysdig, put the need plainly: "What I didn't have was an ability to integrate and access all the different data points. Spotlight brings it all together and spotlights what needs to happen next."
The same components serve the wider team. Sales inspects and forecasts deals on them. RevOps sees velocity and slippage patterns across the pipeline, Value teams see which deals lack discovery and value depth, and Customer Success applies the same score to renewals.
Reading a score that just dropped
A drop is only useful if you can tell what kind of drop it is. The same ten points can mean very different things for the next call.
Engagement fell
Check who stopped showing up. If the champion is still active and the economic buyer has gone quiet, the next step is getting the champion to bring the buyer back. If everyone went quiet at once, look for an event on the buyer's side, such as a reorg or a budget freeze.
A qualification letter lost its support
Sometimes the score drops because the evidence behind a letter weakened: the decision process the buyer described in discovery changed on the last call, or the metric they cared about no longer comes up. That is a discovery task for the rep, and the score has pointed them to the exact letter.
The deal aged
When the only component that moved is time on stage, nothing new went wrong. The deal has stayed in place long enough to carry more risk than its stage suggests, and a direct question to the buyer about timing usually settles it.
Auditing the deal health score you have now
Pick three deals that slipped last quarter. Pull their health scores from four weeks before the slip.
Break each score into inputs. If the platform cannot do this, note it; that is your first finding.
Mark each input as activity or evidence. Email counts are activity. A buyer naming a decision date on a call is evidence.
Check for missing-person penalties. See whether a deal with no engaged economic buyer still scored green.
Look at time on stage. Compare days in stage against your historical average and see whether the score reacted.
Ask a rep what they would have flagged. Reps often saw the risk first. A score that ignores what they knew has a gap.
Rebuild one slipped deal's score
Take the clearest slip from the audit and rebuild its score by hand from the conversations: who attended, what the buyer committed to, which MEDDPICC letters had evidence behind them. Compare it with what the platform showed at the time.
If the hand-built version would have warned you weeks earlier, the gap is in the inputs. A Spotlight.ai demo can score the same deal on its components so you can see the difference side by side.
FAQs About Deal Health and Risk Scoring
What is a deal health score?
A single measure of how likely an open deal is to close as forecast, built from signals such as activity, qualification, stakeholder engagement and slippage. Its value depends on which inputs it uses.
What platforms track deal health and risk?
Forecasting and RevOps platforms, conversation intelligence platforms and CRMs all offer health or risk indicators. Compare them on whether the score uses buyer evidence and whether you can see which input moved it.
Why do deals with high health scores still slip?
Many scores weight activity heavily, so a busy deal looks healthy even when the economic buyer has gone quiet. Activity measures motion and says little about who is engaged.
What should a deal health score include?
Qualification status backed by evidence, stakeholder sentiment and influence, penalties for a missing champion or economic buyer, time on stage and slippage history, alongside activity.
Can a deal health score be inspected?
It should be. You should be able to open any deal's score, see each component's value and trace it to the conversation or calculation behind it.
How does the Spotlight Score work?
Each component is scored from minus 1 to plus 1 and combined as a weighted adjustment to the deal's stage probability. It covers qualification, stakeholders, velocity, time on stage, win and loss patterns and slippage.
Comments