The forecast everyone pretended to believe.
How rewriting the incentive structure turned a $5M end-of-quarter forecast variance into $167K — calculated 90 days before the quarter closed.
In management consulting, forecasting is treated as a problem of nature. Senior consultants who've been closing million-dollar engagements since before Salesforce existed don't log pipeline with discipline, and getting them to is considered a lost cause — so you triangulate, add a buffer, and hope. At this consulting division inside a larger EPC firm, the forecast being off was treated as a law of physics. The real story underneath was different: the CRM had ballooned to 200+ fields through acquisitions and layered-on sales methodologies (blue sheets, green sheets), and senior consultants said openly that filling it out gave them nothing. There were no consequences for vagueness — and vagueness, it turned out, was protecting an arrangement. A senior consultant carrying a $3M quota with a $6M deal that kept slipping, kids in college, retirement close, and all the leverage in their pocket has no reason to put specificity into the CRM. Specificity creates accountability. Vagueness protects the deal. The company wasn't being failed by lazy salespeople — it was being held hostage by rational actors. That isn't a technology problem. That's a tell.
- 01Made the CRM usable. Stripped fields, stages, and anything not directly required to move a deal from lead to close. Table stakes — but it had to come first.
- 02Paid reps to forecast better. A meaningful portion of comp was tied to hitting commits within a defined variance of forecasted value. The only way to qualify was to enter deals early and maintain them. Living outside the system stopped being free.
- 03Held sales leaders accountable for their reps' CRM health. Managers were drilled weekly on pipeline health. Any deal mentioned that wasn't in the system triggered an immediate question from the COO or CEO. No air cover for working off-system.
- 04Made pipeline hygiene a public practice. Wins celebrated publicly. Overdue opportunities named publicly — not punished, named. Deals not in CRM were treated as fantasy. Culture shifted from individual opacity to collective accountability.
The division became the gold standard for pipeline health across the entire company. End-of-quarter forecast variance dropped from $5M to $167K in a $150M business — and that $167K was calculated 90 days before the quarter closed. At the beginning of the quarter, leadership could predict where it would land within $167K. Initiatives moved faster because the numbers were trustworthy. The lesson: your CRM isn't broken and your data isn't the cause of your problems — it's the result. The question is whether the people responsible for the data have any real, financial, consequential reason to care that it's accurate. If they don't, no tool — AI or otherwise — will fix it. Show me the behavior, I'll show you the incentive. For humans and AI.