B2B sales forecasting is the process of predicting how much revenue the sales team will close in a defined future period -- typically the current quarter or the next 30, 60, and 90 days. Accurate forecasting matters for two reasons: it allows the business to plan resources, headcount, and cash requirements (inaccurate forecasts create over-hiring and over-spending in good periods, and panic cuts in bad periods); and it is a diagnostic tool for pipeline health (a persistently low forecast vs. actuals gap signals either over-optimistic qualification or a pipeline generation problem). The method used to build the forecast determines how accurate it can be and what behaviours it incentivises in the team.
Stage-weighted forecasting
Stage-weighted forecasting assigns a probability percentage to each pipeline stage and multiplies the deal value by that probability to get the expected value. Example: Discovery = 10%, Proposal = 30%, Negotiation = 60%, Verbal Commit = 90%. A 1 Cr deal in Proposal stage contributes 30L to the forecast. The weighted pipeline is summed to produce a total forecast number. Advantage: simple, consistent, visible to everyone. Disadvantage: stage probability is a historical average, not a deal-specific prediction -- a 1 Cr deal in Proposal stage at an account where the champion is strong and the decision is next week should forecast very differently from a 1 Cr deal in Proposal where the champion is new, the economic buyer has not been engaged, and the timeline is "end of quarter." Stage-weighted forecasting is the starting point for most teams but should not be the only method used.
Historical rate forecasting
Historical rate forecasting uses the team's actual conversion rates at each stage to project forward from current pipeline. If the team historically converts 25% of deals from Discovery to Closed Won, 100L of pipeline in Discovery stage should produce 25L in closed revenue by a given date. This method is more accurate than stage-weighted when applied to large enough pipeline samples (individual rep forecasts have too much variance; team-level forecasts with 30+ deals in pipeline are more predictive). Historical rate forecasting breaks down when market conditions change, when the pipeline composition shifts (new segments, new product lines), or when the team composition changes significantly.
Commit and best-case forecasting
In the commit/best-case model, each rep categorises their deals into three buckets: Commit (the rep is highly confident this deal will close in the period -- they would be surprised if it did not), Best Case (the deal could close if conditions go well, but there is meaningful uncertainty), and Pipeline (this deal is not expected to close in the current period). The rep's forecast = committed deals + a percentage of best-case deals. Managers then roll up rep forecasts with a haircut to account for the fact that reps systematically over-forecast. This model is intuitive and creates clear accountability (a rep who commits a deal is professionally accountable for closing it) but is highly dependent on rep sandbagger/bluebird culture -- reps who systematically over-commit or under-commit make the model unreliable.
AI-assisted and opportunity scoring
AI-assisted forecasting uses machine learning models trained on historical CRM data to predict the probability of each individual deal closing, based on deal attributes: deal age, stage, engagement signals (email opens, meeting activity, stakeholder contacts engaged), company size, rep tenure, product, and competitive situation. Tools like Clari, Gong, and Salesforce Einstein apply these models at the deal level and roll them up into team forecasts. AI forecasting is the most accurate method for teams with enough historical data (typically 500+ closed deals) and clean CRM data, but it is a black box to reps and managers, and requires ongoing model maintenance as business conditions change. It should complement rep-submitted commit forecasts rather than replace them.
Frequently asked questions
- What are the main B2B sales forecasting methods?
- The main B2B sales forecasting methods: (1) Stage-weighted: assign a probability to each pipeline stage and multiply by deal value; simple and consistent but averages across individual deal quality. (2) Historical rate: use actual stage-to-stage conversion rates from past quarters to project forward from current pipeline; more accurate for large pipeline samples. (3) Commit/best-case: reps categorise deals as Commit (high confidence), Best Case (could close), or Pipeline (not this period) and submit a rep-level forecast; creates accountability but depends on rep calibration. (4) Opportunity scoring: use deal-level signals (stage, engagement, age, stakeholder count) to score each deal individually and roll up; more accurate than stage-only methods for individual deals. (5) AI-assisted: machine learning models trained on historical CRM data to predict deal-level close probability; most accurate for teams with clean data and sufficient deal history, but requires ongoing maintenance.
- How do you improve B2B sales forecast accuracy?
- To improve B2B sales forecast accuracy: (1) Improve pipeline qualification -- inaccurate forecasts are usually a pipeline quality problem, not a forecasting method problem; deals that should not be in the pipeline inflate the forecast; (2) Use multiple methods and compare them -- a forecast produced by three methods (stage-weighted, commit-based, historical rate) is more accurate than any single method; when the three methods diverge significantly, investigate why; (3) Track forecast vs. actuals weekly and root-cause gaps -- is the miss from deals slipping (timing issue) or deals being lost (qualification issue)? The root cause changes the fix; (4) Apply consistent MEDDIC or equivalent qualification to ensure that deals in each pipeline stage genuinely meet the stage criteria; stage-inflation (moving deals to advanced stages before they qualify) is the most common forecasting accuracy problem; (5) Train reps on commit calibration -- help reps understand the difference between what they hope will close and what they are genuinely confident will close.
- What is a commit in B2B sales forecasting?
- A commit (or "upside commit" in some organisations) is a deal that a sales rep formally categorises as highly likely to close in the current forecast period. When a rep commits a deal, they are professionally accountable for closing it: a committed deal that does not close is a forecast miss and requires explanation. The commit category sits above "best case" (deals that could close but have meaningful uncertainty) and above "pipeline" (deals not expected this period). The accuracy of the commit category -- how often committed deals actually close -- is one of the most important measures of a rep's forecast quality. High-performing reps who over-commit (putting too many deals in commit) are actually a bigger forecasting problem than reps who under-commit, because they inflate the forecast and create false confidence in the revenue plan.
Keep reading
- Sales forecasting: what it is and how to forecast B2B revenue
- B2B revenue forecast: how to build a B2B revenue forecast
- B2B pipeline management: how to manage your B2B sales pipeline
- B2B pipeline health: how to assess whether your sales pipeline is real
- What is MEDDIC? The enterprise sales qualification framework