
Pipeline & Qualification · Module 3 of 4
Walk or Work It: When to Kill a Deal You've Already Bled For
You've poured 60 hours into Meridian — the CRO's favorite $180k logo — and it's stalled at a 15% chance. Every instinct says "we're too close to quit." This is the mid-quarter reckoning: four of Arclight's deals have stalled, each with real hours already sunk, and for each you decide continue or kill — where killing frees the remaining hours to redeploy to live pipeline at about $2k an hour. You're the senior AE, and the rule is forward-only: a sunk cost is gone whether you continue or kill, so it belongs in neither number; continue a deal only if finishing it (its win-probability times its value) beats what its remaining hours would earn redeployed. Meridian proves it to the dollar — finishing is worth $27k while redeploying its 50 hours is worth $100k, so continuing destroys $73k, and it would need a 56% win to justify staying against its honest 15%; the 60 sunk hours, the loudest argument in the room, never enter the math. But the rule isn't a license to quit — it cuts both ways. Kepler ($66k to finish vs $40k redeployed, +$26k) and Cardinal ($80k vs $50k, +$30k) quietly clear the bar, and the cynic who "kills all the stalled junk" wrongly culls them too. The disciplined triage — continue Kepler and Cardinal, kill Meridian and Tidewater — is worth $276k, beating both reflexes: never-quit (continue all four) lands at $200k, the worst, and kill-everything at $220k. Then the same forward logic re-decides when the world changes: new information lifts Meridian to 60% and you revive the deal you correctly killed ($108k > $100k), and an hours squeeze to 40 makes the two good continues compete, keeping the higher-surplus Cardinal at $110k. Fight for forward yield, not for logos or for what you've already spent. This is the third of the four Pipeline & Qualification modules — qualify by expected value per hour, forecast the number honestly, and walk away from what no longer pays, so the quarter's hours land on the deals that actually pay. One module remains: how much pipeline to build when the number isn't covered.
0 of 56 items completed0%~110 min remaining
Material · 56 items