Forecast

Origin & evolution

Forecasting began as the front-line gut call: a rep eyeballing the pipeline and naming a number. The structuring discipline came from CRM, which formalised forecast categories. The model most teams now use, codified in Salesforce's default categories, sorts open deals into Commit (you will stake your name on these), Best Case (plausible upside), and Pipeline (everything else still in play). The categories separate confidence levels so a single blended number does not hide the difference between a deal that is signed-bar-the-signature and one that is a hopeful first call.

A parallel technique, the weighted pipeline, multiplies each deal's value by a stage-based probability and sums the result. It is simple and widely used, and its weakness is well documented: it ignores deal quality, rep behaviour, and market conditions, so one large deal at the wrong probability distorts the whole figure. Mature practice keeps both lenses, unweighted coverage to diagnose whether enough is in flight, category-based judgement to make the quarterly call.

Two human biases plague the number. Sandbagging reps deliberately call low and beat it, consistently delivering above 120% of forecast; happy-ears reps believe every prospect and call high. Accurate reps land within roughly 90 to 110% of actual bookings, and tracking per-rep accuracy is how a manager learns whose number to trust.

How it works in practice

A regional team carries £2.1m of open pipeline. The manager does not report £2.1m. She sorts it: £620k in Commit, £340k in Best Case, the rest in Pipeline. She calls £680k, slightly above Commit, having added one Best Case deal she has independently verified with the economic buyer. One rep's number she discounts on sight, his last three quarters came in at 130% of his call, classic sandbagging, so his real commit is higher than he admits. The quarter closes at £695k. Her forecast accuracy lands at 98%, and that track record is what makes finance plan against her number without second-guessing it.

Forecast vs. its neighbours

• Pipeline is the full inventory of open deals. The forecast is the disciplined subset the team will actually commit to, after judgement and category sorting. The pipeline is everything possible; the forecast is what is probable.
• Revenue stream💵Revenue StreamBusiness ModelA source of revenueView reference → is the realised, recurring money a product brings in. The forecast predicts future closes; the revenue stream is the booked result the forecast was trying to anticipate.
• Metric🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference → is a measured fact about the business. The forecast is a prediction, and forecast accuracy is itself a metric: the gap between what was called and what arrived.

In the graph

In the Unified Product Graph, forecast🔮ForecastSales & RevenueA revenue forecast sits in the sales domain of the Business, GTM & Growth region. It links to what it is about and what it produces through three edges: product_forecasted_via_forecastProductforecasted viaForecasthierarchy ties the prediction to the product, pipeline_sales_projected_via_forecastPipeline Salesprojected viaForecasthierarchy derives it from the live pipeline, and forecast_predicts_revenue_streamForecastpredictsRevenue Streamcross-domain connects the call to the revenue it anticipates. Separating the forecast from both the pipeline it reads and the revenue it predicts is what lets the graph measure accuracy: the predicted figure and the realised figure are distinct, connected nodes, so the distance between them is queryable.