Sales forecasting in HubSpot: making the number believable

Most forecasts fail for one reason: the deal data behind them is optimistic. Fix the inputs before arguing about the model.

Start with exit criteria

Every stage needs an objective condition to leave it — a completed technical validation, a received requirements document, a confirmed budget holder. Without exit criteria, stage probability is a feeling with a percentage attached.

Use three views, not one

Weighted pipeline (probability by stage), commit forecast (what reps commit to), and best case. Leadership decisions use the commit view; capacity planning uses weighted. Reporting one number hides the risk.

Keep close dates honest

A close date that moves every week is a data quality signal, not a scheduling issue. Report on close-date pushes per rep and per deal. Three pushes usually means the deal was never qualified.

Segment before you trust the average

Forecast accuracy improves dramatically when you split by segment: new business vs expansion, deal size band, source, and country. A single blended conversion rate hides two very different businesses.

Add velocity to the picture

Pipeline value alone cannot tell you whether the quarter lands. Combine coverage ratio (pipeline vs target), average cycle time and stage conversion. If cycle time exceeds days remaining, the deal is next quarter regardless of the rep's optimism.

Automate hygiene, do not nag

Workflows should flag deals with no activity in 14 days, missing amounts, past close dates or stages without required properties. Send the flags to owners and managers, weekly, in one digest.

Review the forecast against reality

Every month, compare the forecast made 30 days earlier with what closed. Track the error, by rep and by segment. Forecasting is a skill that improves only when the miss is measured.