Overview
Every founder has lived this month: the forecast said 300K, the month closed at 140K, and the explanations all sounded reasonable. Deals slipped. A whale went quiet. Two "sure things" evaporated. The forecast was not unlucky. It was built from the wrong raw material: rep optimism, weighted by stage percentages that mean nothing.
A trustworthy forecast is built from three inputs, all of which you already have if you run the systems in this knowledge base.
Input 1: Expected close dates, harvested on calls
The traditional forecast asks reps to guess when deals will close. The disciplined floor does not guess, because the follow up workflow harvests the date from the prospect on the call: when they say "I'll decide by Friday," the rep logs Friday as the expected close date and sets the day-before touch. A forecast built from prospect-stated dates, each with a scheduled day-before check-in that surfaces slippage 24 hours early, is a different species from a forecast built from rep vibes.
The rule that makes this work is the same rule that fixes pipeline reviews: no next step and no date means the deal is not in the forecast at all. It is a lead in a nurture sequence. This rule alone typically deletes a third of a fictional pipeline, and everything it deletes was never going to close this month anyway.
Input 2: Behavior-based staging, not stage-percentage theater
Standard CRM forecasting multiplies deal value by a stage percentage, proposal sent equals 60 percent, and sums it. The percentages are decorative. A proposal sent to a prospect who has gone silent for two weeks is not 60 percent of anything.
Stage deals by prospect behavior instead, the hot warm cold definitions the floor already uses: hot means committed or near-committed with a date, warm means has pricing but uncommitted, cold means a month-plus timeline. Then forecast in two lines only. Commit: hot deals with expected close dates inside the month. Upside: warm deals with dates inside the month. Cold deals and dateless deals forecast at zero, whatever their size, because behavior says zero.
Input 3: Your floor's own conversion history
Now calibrate with the only percentages that mean anything: yours. Over the trailing quarter, what fraction of hot-with-date deals actually closed inside their month? What fraction of warm? Those two numbers, hot conversion and warm conversion, are your floor's real weights. Multiply commit by the first, upside by the second, add new business you can expect from this month's calendar at historical close rates, and you have a forecast with error bars you measured instead of invented.
Most floors discover their hot conversion runs 60 to 80 percent and warm runs 15 to 30. Whatever yours are, they are yours, and they tighten every quarter you keep score.
The weekly forecast rhythm
The forecast is not a monthly document, it is a weekly output of the pipeline review. Every week: deals with close dates inside seven days get plans confirmed, slipped dates get moved honestly instead of silently, zombie deals get killed, and the two-line forecast gets recomputed in ten minutes. A forecast touched weekly degrades gracefully. A forecast built on the first and checked on the thirtieth just detonates.
And hold the honesty at the top: when a manager massages the number upward to look good in the founder meeting, the entire machine below is wasted. The forecast's job is to be right, not reassuring, and a floor learns within one quarter which one leadership actually rewards.
What this buys the founder
Cash planning that works, hiring decisions timed to real revenue, and most underrated: an early warning system. When the commit line is thin on the 5th, you know on the 5th, while there is still a month of calendar to fill and follow up sequences to run, instead of discovering it in the closing week. A trustworthy forecast does not just predict the month. It gives you time to change it.
Summary
Forecast only deals with prospect-stated close dates and a booked next step. Stage by behavior, hot and warm, and forecast in two lines. Weight by your own trailing conversion history, recompute weekly inside the pipeline review, and reward accuracy over optimism. The forecast stops being fiction the same week the pipeline does.
Frequently asked questions
What accuracy is realistic?
Floors running this land within 10 to 15 percent of forecast most months, and the misses come with early warning. Vibes-based forecasts routinely miss by 40 plus, in one direction.
How do long sales cycles change this?
Nothing structural. Dates are further out, the day-before touches still fire, and the trailing conversion weights simply get computed over a longer window.
Should reps see the floor forecast?
Yes. Reps who see how commit is built stop inflating stages, because inflation now has a visible owner and a visible cost.