Every revenue tool now has a red flag. A deal goes quiet, a champion stops replying, a close date slips twice, and a risk score turns amber. Those alerts are useful, and you should keep them. But a risk alert answers one question: which live deal needs attention now? It was never built to answer a different one: why is our win rate what it is?
What a risk alert is built to do
A risk model scores one open opportunity against a set of signals the vendor chose: no next step booked, a single contact, no reply in fourteen days, a stage held too long. It is tuned to catch deals before they slip, and it is judged on whether the flagged deals really did slip. Inside a forecast call that is exactly the right job.
Why alerts don't add up to an explanation
- They see the deal while it is open. The decision that lost a deal is often made weeks before the signals appear. By the time a thread goes quiet, the cause is in the past.
- They count what was instrumented. An alert can only fire on signals someone thought to measure. A scope cut agreed in Slack, or a budget figure said aloud on a call, never becomes a signal.
- They never compare with the wins. A flag says this deal looks risky. It does not say whether the same pattern appears just as often in the deals you won, which is the only test that tells you whether it explains anything.
- Nobody counts them across a period. A year of alerts is a year of separate notifications. It is not a denominator.
What explaining a win rate needs
Explanation needs every deal you have already worked, joined across the systems that recorded it, counted, and compared with the wins. That is the method we call win-rate forensics: each finding carries a denominator, evidence you can put on screen, a counterfactual from the won deals and a named owner. It shows which behaviors across your GTM process are associated with winning and losing.
An example
In Case 01, a cloud storage company had 14 large deals that were cut down to fit a stated budget. None was won. A risk model would most likely have flagged those deals only when their email threads went quiet, weeks after the scope was cut on a pricing call. The pattern worth acting on sat earlier, in how budget was qualified, and it only became visible when all 62 qualifying deals were counted together.
Alerts are for this week's pipeline. Forensics is for next quarter's process.
Use both
Keep the alerts in the forecast call, where they belong. When the win rate moves, or two leaders disagree about why, the answer comes from counting across every deal, not from adding up the flags.
About Rev-Lens. Rev-Lens is a founder-delivered go-to-market intelligence practice. In an eight-week engagement it joins a company's CRM, recorded calls, email, internal chat and intent data at the opportunity level, inside the company's own infrastructure, and returns win-rate findings by week five, each with a denominator, the evidence behind it and a named owner. The company keeps the joined dataset and the agents, and its team is trained to run them.
Keep reading: the white paper Reach Is Not Analysis, or how Rev-Lens compares to conversation intelligence, forecasting and win-loss research.