Sales Forecasting: How to Build a Quarterly Number That Lands Within ±10%

A quarterly number accurate to within ±10% doesn't come from summing opportunities multiplied by their stage probability. It comes from splitting the pipeline into three layers — Commit, Best Case, Gap — and deriving probability from your own historical win rate rather than a default table someone shipped with the software. And no calculation means anything until four fields are clean: close date, stage, amount, owner.

Why Your Number Drifts: Three Mistakes That Repeat Every Quarter

One: fixed probability percentages hard-coded into the system a year ago that nobody has revisited. If your actual win rate from "Proposal Sent" is 35% and you're calculating it at 60%, you're inflating every quarter by a third.

Two: close dates that get pushed forward month after month. An opportunity whose date has slipped three times is not a this-quarter opportunity, and leaving it in the calculation is the single largest source of variance.

Three: confusing forecast with target. The rep quotes a number that keeps the manager happy, the manager rounds it up to keep leadership happy, and finance receives a figure nobody actually said.

The Three Layers: Commit, Best Case, Gap

• Commit: deals where the decision has genuinely been made and only execution remains — an announced award, a PO being issued, a contract sitting with the signatory. Counted at roughly 100%.
• Best Case: active opportunities with a close date inside the quarter, an authorised point of contact, and documented activity within the last 14 days. Counted at the historical win rate for their stage.
• Gap: the difference between the two layers combined and the quarterly target. This isn't a number you wish for — it's a task, translated into how many new opportunities must be created this month.

Take your closed deals from the last four quarters and, for each stage, calculate how many opportunities entered it and how many of those closed won. The resulting percentage is your real probability. Run it separately for each segment when the dynamics differ — public sector behaves differently from private, both in win rate and in cycle length. Refresh the figures once per quarter: no more often, no less.

Six Hygiene Rules Before Any Calculation

1. No opportunity without a specific close date.
2. Close date pushed twice or more → it drops from Commit to Best Case.
3. No documented activity in 21 days → it leaves the calculation until it's updated.
4. Amount is recorded at current-quarter value only, not the full multi-year contract value.
5. Every opportunity has one named owner.
6. Stage changes on an objective event (proposal sent, meeting held), never on a feeling.

A Worked Example: Forecasting a Full Quarter Step by Step

A hypothetical company with a quarterly target of SAR 2 million:

The gap is SAR 520,000. At a blended win rate of 25%, that means creating roughly SAR 2 million in new opportunities — a marketing and prospecting task that starts today, not in the final month of the quarter.

The Weekly Forecast Review: What to Ask, What to Move

A 30-minute meeting that discusses movement only: what entered Commit, what left it and why, and which close dates changed. Three questions for every deal in Commit: who signs? What's the next step and its date? What's blocking it? If even one answer is missing, the deal drops a layer.

Measuring Forecast Accuracy: How You Know You've Hit ±10%

At the end of every quarter, record two figures: the number you forecast in week one, and the number you actually landed. Variance = (actual − forecast) ÷ forecast. Track it across four consecutive quarters. If the variance is consistently positive, your weights are too conservative; if it's consistently negative, the hygiene rules aren't being enforced. The difference between 30% and 10% isn't predictive genius — it's discipline in four fields and a measurement you repeat.