The five quote-to-cash metrics that matter (and three that don't)
Deal desks generate a lot of numbers and very little measurement. These five metrics predict revenue outcomes; each comes with a note on what to instrument so it isn't estimated by feel.
1. Quote turnaround time
Opportunity marked "needs quote" → quote sent. The single best predictor of win rate you control directly — in competitive deals, first credible number anchors the negotiation. Instrument: timestamp quote creation and first send; report the median, not the mean (one gnarly enterprise quote shouldn't hide ten fast ones).
2. Approval latency by tier
Approval requested → decision, split by threshold tier. If your 10–20% tier takes two days, reps will learn to quote 9.9%. Instrument: log request and decision events per step; alert on anything pending >24h.
3. Version churn
Versions per closed quote. Two to three is a healthy negotiation. Six or more means pricing wasn't credible to begin with — usually a catalog or rules problem, not a rep problem. Instrument: free if your CPQ versions automatically; impossible if it overwrites.
4. Attach rate
Share of quotes including attach targets (support plans, onboarding, add-ons). This is where "AI recommends bundles" either earns money or doesn't. Instrument: tag attach SKUs; compare rates with and without recommendations shown.
5. Discount leakage
Gap between guideline price and realized price, aggregated. Covered in depth in the governance post — the one metric that converts directly to a dollar recovery number.
Three metrics to skip
- Quotes generated per rep. Volume without outcome; incentivizes spam quoting.
- "CPQ adoption %." If reps route around the tool, measure why (usually approval latency), not the routing itself.
- Average deal size in isolation. Moves with mix, not with anything your deal desk did this quarter. Watch realized-vs-guideline price instead.
Start with two
If you instrument nothing else this quarter: median quote turnaround and approval latency by tier. Both are pure event logging, both have obvious targets, and both change rep behavior the week you publish the dashboard.