Ask most sales leaders why the forecast missed and you'll get a story about one or two deals that slipped. Ask why it misses every quarter and the room goes quiet.
Forecasting isn't hard because the future is unknowable. It's hard because most forecasts are built from optimism, stage percentages nobody validated, and close dates reps invented to fit the quarter.
The three methods, honestly
Stage-weighted. Assign a probability to each stage, multiply, sum. Fast, and only as good as probabilities derived from your actual closed history. Most teams use numbers someone made up in a planning meeting years ago — 20/40/60/80 — and then wonder why the output is wrong.
Rep commit. Ask reps what will close. Captures real deal knowledge no model has. Also captures optimism, quota pressure, and reluctance to admit a deal is dying.
Evidence-based. Forecast only deals passing explicit checks: confirmed close date from the buyer, named economic buyer engaged, documented decision process, no unresolved blockers. Slower, considerably more accurate.
The practical answer is a hybrid: stage-weighted as the baseline, evidence checks as the gate, rep commit as an input you interrogate rather than accept.
Fix your stage probabilities first
If you take one thing from this piece: derive your stage probabilities from your own closed history, don't inherit them.
Pull two years of closed deals. For each stage, calculate what share of deals reaching it eventually closed won. That's your real probability. It's almost never the round numbers in your CRM, and the gap between the two is a large share of your forecast error.
Recalculate quarterly, and segment it. Enterprise deals at "proposal sent" and SMB deals at "proposal sent" are different animals wearing the same label.
Categories that mean something
The categories only work if the definitions are enforced.
| Category | Definition | Treat as |
|---|---|---|
| Commit | You'd bet the quarter on it. Buyer confirmed timing and budget; only execution remains | ~90% |
| Best case | Could close if things break your way | ~50% |
| Pipeline | Real opportunity, this quarter unlikely | ~20% |
| Omitted | Not happening this period | 0% |
A commit that slips twice was never a commit. The moment "commit" becomes a hope with a confident tone, the whole forecast stops meaning anything.
Protect the definition ruthlessly. It's the only part of forecasting that's genuinely cultural rather than analytical.
The questions that test a forecast
In a forecast review, don't ask "will it close?" Ask:
- What's the close date, and who told you — you or them?
- Have you met the person who signs?
- What's left between now and signature?
- How long does their legal review usually take?
- What's the single most likely reason this slips?
- What did the buyer do in the last week?
Question 6 is the fastest diagnostic in the set. A commit deal with no buyer-side activity in two weeks is not a commit.
Measure your own accuracy
Almost nobody does this, and it's the highest-leverage habit available.
Each quarter, record what you forecast at week 1, week 6, and week 11, and what actually closed. Track the error. Within three quarters you'll see your own pattern — most teams have a stable, predictable bias, and knowing yours is worth more than any methodology change.
Common patterns:
- Consistently 15–20% over — deals slip and reps over-commit; tighten the commit definition
- Accurate at week 11, wildly wrong at week 1 — early-stage probabilities are miscalibrated
- One rep always wrong — a coaching conversation, not a model problem
- Accurate on total, wrong on which deals — your process is fine; your deal-level insight isn't
The failure modes
Sandbagging. Reps under-forecast to beat the number. Makes forecasts useless in the other direction, and it's a compensation design problem rather than a character problem.
Happy ears. A good call feels like progress. It isn't, unless something changed on the buyer's side.
Hero deals. One enormous deal carrying the quarter. It slips and everything misses. Watch concentration risk explicitly.
Quarter-end compression. If most deals close in the last two weeks, you don't have a forecasting problem, you have a pipeline distribution problem — and the compression is what buyers exploit for discounts.
Zombie deals. Deals that never die and never close, sitting in late stages inflating coverage for quarters at a time.
A workable cadence
Weekly: reps update deals; commits get inspected against evidence, not confidence.
Monthly: roll up by segment; compare against the same point in prior quarters.
Quarterly: recalculate stage probabilities from closed history; review your forecast accuracy; adjust the process based on the error pattern rather than the anecdote.
The point isn't a perfect forecast. It's a forecast wrong by less each quarter, and known to be wrong in a consistent direction — which is enough to run a business on.
Frequently asked questions
What is the most accurate sales forecasting method?
For most teams, a hybrid: stage-weighted pipeline as the baseline, adjusted by explicit evidence checks — is there a confirmed close date, a named economic buyer, and a documented decision process. Pure stage-percentage forecasting is only as good as percentages you've actually validated against closed history, which most teams never do.
Why is my sales forecast always wrong?
Usually one of three causes: stage probabilities that were guessed rather than derived from history, reps forecasting on optimism instead of buyer evidence, and deals sitting in late stages long past a realistic close date. Track your forecast error each quarter and the pattern behind your particular miss becomes obvious quickly.
What's the difference between commit and best case?
Commit means you would bet your quarter on it — the buyer has confirmed timing and budget, and only execution remains. Best case means it could close with everything breaking your way. Blurring them is how forecast credibility dies: a commit that slips twice was never a commit.
Put this playbook to work
Twin-Sales reads your market, finds the accounts worth pursuing, writes outreach in your voice, and runs your pipeline from first touch to closed deal.
Try Twin-Sales free →