# Sales Forecasting — Standard Operating Procedure > Source: https://b2bprocess.com/sales-forecasting > Last updated: 2026-07-08. Adapt owners, tools, and thresholds to your organization. ## 1. Purpose Sales forecasting is the recurring process of predicting how much revenue will close in a given period — typically rolled up weekly from rep to manager to leadership using defined categories (commit, best case, pipeline), validated against pipeline data and deal inspection, and scored afterward against what actually happened. It converts a pile of open opportunities into a number the company can plan hiring, spend, and investor communication around. ## 2. Scope & prerequisites Formalize forecasting once there are multiple sellers and a leadership team making decisions on the number — effectively from the second AE onward. Prerequisites: defined opportunity stages with exit criteria, enforced hygiene on amounts and close dates, and a CRM culture where the pipeline is the truth rather than a compliance chore. ## 3. Roles & responsibilities | Role | Responsibility | | --- | --- | | Sales rep / AE | Owns per-deal calls and honest category assignments backed by evidence. | | Front-line sales manager | Inspects deals, calibrates judgment, owns the team number. | | CRO / Sales leadership | Owns the company number, its communication, and the culture that makes honesty safe. | | Revenue Operations | Owns hygiene, snapshots, category governance, triangulation models, and accuracy scoring. | | Finance (FP&A) | Consumes the forecast for planning; reconciles bookings actuals; co-owns board presentation. | ## 4. Procedure ### Step 1: Define forecast categories with teeth **Owner:** Sales leadership + RevOps Define each category behaviorally, not vibes-ly: Commit = 'I will resign over this number' territory — paper in legal, signature process known, date confirmed by the buyer. Best case = real upside with a named, plausible path. Pipeline = qualified but unproven. Write examples and anti-examples; calibrate in the first month of manager reviews. - [ ] Document category definitions with objective evidence requirements - [ ] Map categories to stages loosely, not mechanically — category is judgment, stage is fact - [ ] Train reps and managers on the same cases ### Step 2: Enforce the hygiene the forecast stands on **Owner:** Revenue Operations Weekly automated checks: opportunities past their close date, stages stale beyond thresholds, missing amounts or next steps, deals pushed more than twice. Hygiene exceptions are fixed before the forecast call, so the call spends its time on judgment rather than data archaeology. ### Step 3: Snapshot everything, weekly **Owner:** Revenue Operations Freeze pipeline and forecast state at the same time each week. Snapshots make the invisible visible: slippage, category migration, late-quarter hockey sticks, and each forecaster's historical accuracy. Without snapshots the forecast has no memory and no accountability. - [ ] Automated weekly snapshot of every open opportunity and its category - [ ] Report week-over-week movement (new, advanced, slipped, lost) by team - [ ] Retain history for accuracy scoring and seasonality analysis ### Step 4: Run the rep-level forecast submission **Owner:** Reps Each rep submits their number by category with per-deal calls before the manager 1:1 — in the tool, not verbally. The act of writing it down against named deals is the first defense against drive-by optimism. ### Step 5: Inspect deals, not spreadsheets, in the manager review **Owner:** Front-line managers The weekly forecast 1:1 pressure-tests the calls with evidence questions: who is the economic buyer and when did we last talk to them? What is the paper process and where is it? Why does the close date say the 28th — whose date is that, ours or theirs? Managers adjust judgments and coach the gaps; this meeting is where forecast accuracy is actually manufactured. - [ ] Standard inspection questions per category claim - [ ] Downgrade deals that fail evidence tests — visibly and consistently - [ ] Log risks and next actions on the opportunity, not in private notes ### Step 6: Roll up with judgment layered on data **Owner:** Sales leadership + RevOps Leadership assembles the company number from the manager roll-ups, triangulated against independent signals: historical stage-conversion rates applied to current pipeline, category-accuracy history by team, pipeline coverage and age, and (where used) AI/statistical forecasts. Divergence between the human number and the model number is the most useful conversation of the week. ### Step 7: Communicate one number with its assumptions **Owner:** CRO / Sales leadership The forecast that leaves the sales org states the number, the range, the key swing deals, and the assumptions that would move it. Finance and the board get consistency: same definitions, same cadence, same format, quarter after quarter. ### Step 8: Score accuracy and feed it back **Owner:** Revenue Operations After each period: forecast-vs-actual by rep, manager, and category, at each week of the quarter. Publish it. Persistent sandbaggers and persistent optimists are both accuracy problems with names, and both improve remarkably when their track record is visible. Feed systematic biases into category definitions and coaching. - [ ] Score week-N forecast vs. final actuals for every forecaster - [ ] Review misses: which deals moved, and what evidence was ignored? - [ ] Adjust definitions, stage criteria, or coaching based on patterns ## 5. Metrics to monitor | Metric | Definition | Formula | Target | | --- | --- | --- | --- | | Forecast accuracy | Closeness of the committed forecast to actuals, tracked by week-of-quarter. | |Actual − Forecast| ÷ Forecast | ±10% by mid-quarter; tightening weekly | | Commit conversion rate | Share of commit-category deals that actually close in period — the honesty meter for 'commit'. | Commit deals won in period ÷ commit deals | > 85–90% | | Slip rate | Share of forecasted deals whose close date moves out of the period. | Slipped deals ÷ forecasted deals | < 20%; watch repeat slippers | | Pipeline coverage | Open qualified pipeline vs. remaining target — context for whether the forecast is even achievable. | Open pipeline ÷ remaining quota | 3–4×, calibrated to actual win rates | | Hygiene compliance | Share of open opportunities passing all hygiene checks at snapshot time. | Clean opportunities ÷ open opportunities | > 90% | | Forecast variance by forecaster | Each rep's and manager's systematic bias (sandbag vs. optimism), from scored history. | Mean signed error per forecaster over trailing quarters | published; trending toward zero | ## 6. Known failure modes | Failure | Symptom | Corrective action | | --- | --- | --- | | Sandbagging as culture | Teams beat forecast by 30% every quarter; leadership can't plan capacity; the board discounts every number. | Score and publish signed error, not just misses; celebrate accuracy, not overperformance against a fiction; separate stretch goals from the forecast. | | Happy ears at scale | Commits built on champion enthusiasm; quarter ends with 'procurement surprised us' three times. | Evidence-based category definitions; manager inspection on paper process, economic buyer, and mutual dates; downgrade without drama. | | Zombie pipeline | Coverage looks healthy; a third of it hasn't had activity in 60 days; conversion models are poisoned. | Automated staleness rules with forced disposition; pipeline age reporting; make closing-lost socially cheap. | | Forecast call as status theater | An hour of reps reading numbers aloud that were already in the tool; no deal gets inspected. | Numbers submitted before the meeting; the meeting inspects the five deals that swing the quarter. | | No snapshots | Nobody can explain what changed since last week; misses are unexplainable; sandbagging is undetectable. | Automated weekly snapshots and movement reporting — the single highest-leverage piece of forecasting infrastructure. | | Mechanical stage-weighted forecasting | The 'forecast' is sum(amount × stage probability) with probabilities nobody has validated in years. | Use weighted pipeline as one triangulation input; validate stage probabilities against cohorted actuals; keep human judgment on the swing deals. | | Punishing honesty | The rep who downgrades a commit gets grilled; everyone learns to hide bad news until week 12. | Leadership treats early bad news as the process working; scoring rewards early accuracy, not late-quarter heroics. | --- This SOP is maintained as part of the B2B process encyclopedia at https://b2bprocess.com. Check the source page for the latest revision.