Every missed quarter produces the same meeting. Someone questions the weighting. Someone proposes new forecast categories. Someone suggests an AI forecasting tool. Almost nobody asks whether the records the forecast is built on were true.
The model faithfully summarizes lies
A forecast is arithmetic on three inputs: which deals are open, what they're worth, and when they'll close. In most orgs, all three inputs are corrupted — not by dishonesty, by neglect. Zombie deals nobody closed out inflate "open." Placeholder amounts corrupt "worth." Close dates pushed quarter after quarter corrupt "when."
Change the weighting model on top of that and you're rearranging decimals on fiction.
The three failures, ranked by damage
1. Coverage built on zombies
The classic: leadership sees 3.2x coverage and plans accordingly. Strip deals with no activity in 30 days and it's 2.1x — which is the difference between "on plan" and a hiring freeze. The worst part is the asymmetry: inflated coverage feels great right up until the quarter ends.
2. Stages without exit criteria
If "Proposal" means "I sent a PDF" to one rep and "verbal yes from the economic buyer" to another, your stage-weighted forecast is averaging two different businesses. Stages need written exit criteria ("demo completed AND economic buyer identified"), visible in Salesforce, enforced in pipeline review.
3. Close dates as wishes
A close date that slipped three times isn't a date — it's a mood. Track close-date history, flag serial slippers, and make the third slip a forced conversation: advance it honestly or close it honestly.
The fix is a system, not a spreadsheet
The teams whose forecasts I'd defend to a board all run the same loop: a nightly sweep that flags mechanical rot (past-due close dates, empty amounts, stale deals), a Monday digest that puts coverage, movement, and slips in front of leadership before standup, and a weekly meeting where every flagged deal gets a decision. No heroics, no quarterly cleanup weekends.
The sweep and the digest are automatable — they're the first two agents in Soleil's catalog, from $300/month flat. The weekly meeting discipline is yours. Want to know how big your zombie problem is right now? That's the free Pipeline Hygiene Audit — read-only, findings in days.
Get the free Pipeline Hygiene Audit
Read-only, no install, no changes — a written findings report on your org within days, whether or not we ever work together.
Book a free callFrequently asked questions
Why is my Salesforce forecast inaccurate?
Almost always because the records underneath it are stale: dead deals padding coverage, close dates that slipped repeatedly, placeholder amounts, and stage definitions no two reps share. The forecast model faithfully summarizes bad records.
What forecast accuracy should a B2B SaaS company expect?
With clean pipeline data and enforced stage exit criteria, in-quarter forecasts within 10% are achievable for most B2B teams. Beyond that, accuracy problems are usually data problems, not methodology problems.