HubSpot Forecast Accuracy: Why the Number Drops at Quarter End

It is the last week of the quarter. Open Sales > Forecast on Monday and the number looks fine. Open it again on Thursday and it has dropped hard, and not a single deal was lost that week.

Nothing happened to your buyers. What happened is that reps finally updated close dates and stages they had left alone for weeks. If your HubSpot forecast accuracy only holds up after a quarter-end cleanup, the problem is upstream of the forecast, and you can catch it with a short weekly check.

What the HubSpot forecast actually reads

HubSpot’s guide to the forecast tool is clear about the default: the forecast is built from deal stages, and each stage stands for how likely a deal is to close. The view shows your goal, closed won, the gap between them and the open pipeline for the period you pick by close date.

In settings you choose which amount the forecast uses: Weighted amount (amount multiplied by deal probability) or Total amount. You can also group deals into forecast categories such as Pipeline, Best case and Commit, and let HubSpot update those categories by workflow when a deal changes stage, as described in the forecast setup article.

Every one of those inputs is typed or moved by a rep. A close date that has already passed keeps the deal in the current period until someone changes it. A deal parked in a late stage keeps that stage’s probability. The forecast is honest arithmetic on whatever the pipeline says.

Why the number moves in the last week

A pattern we see in SaaS teams, and one sales ops people describe in community threads: the pipeline is updated in bursts. Through the quarter, deals sit with optimistic close dates and stages nobody revisits. In the final week, the manager asks for a clean pipeline, reps push dates into next quarter and move stalled deals back or to closed lost, and the forecast drops to what was true all along.

The drop is a late correction of data that was wrong for weeks, which is why a better report rarely fixes it.

A short weekly forecast accuracy check

Run this every Monday, well before quarter end, on the deals in the current forecast period. Filter the deals view by pipeline and close date, then work through four lists.

1. Close date in the past, deal still open. These deals inflate the current period. Owner: the deal’s rep sets a real date or closes the deal by Friday.
2. Deals that skipped stages. A deal that jumped from discovery to contract sent without the middle steps usually carries a probability it has not earned. Owner: the sales manager checks it against your stage exit criteria. If you have none, start with our post on deal stage exit criteria.
3. No activity for longer than the stage normally takes. If a deal has had no logged activity for longer than deals usually spend in its stage, it is stale. Owner: the rep logs a next step or moves it back. The detection rules are in our post on stale deals.
4. Forecast category out of line with the stage. A Commit deal in an early stage, or an old deal still in Best case. Owner: the manager agrees the category with the rep in the one-to-one.

Then look at the result. HubSpot’s forecast accuracy tracking shows how close past forecasts came to what actually closed, so you can judge both the numbers and the process behind them. If accuracy stays poor period after period, the four lists above usually show where the gap comes from.

Who this matters for

Founders and CEOs who get one forecast on Monday and another at quarter end. Heads of sales who own the number and the one-to-ones. RevOps or the HubSpot admin who would run the filters. One practical note: people who use the forecast tool need a Sales or Service seat assigned to them.

Our take

The portal mirrors the process. A forecast built on stages with no exit criteria and close dates nobody owns will move at quarter end no matter which amount or category setup you pick. Adding AI projections or a new dashboard on top of the same data gives you a nicer chart of the same problem.

The fix is cheap in tools and expensive in discipline: one person owns the Monday check, each list has a named owner, and the manager uses the lists in one-to-ones instead of asking “is this still on track”. It takes little time per rep. If nobody on your team will own it every week, the forecast will keep being corrected in the last week, and you should plan around that rather than trust Monday’s number.

FAQ

Why does my HubSpot forecast change at the end of the quarter?

Usually because close dates, stages and categories are updated in a burst during the last week. The forecast reads those fields, so it moves when the cleanup happens. A weekly check moves that correction earlier.

How is forecast accuracy calculated in HubSpot?

HubSpot compares the forecast with the actual closed won amount: the error is the difference divided by the actual, and accuracy is one minus that error, never below zero. Closed won values are based on the deal stage view. You see it in the forecast submission history.

Should the HubSpot forecast use weighted or total amount?

Weighted amount multiplies the deal amount by the deal probability, so it depends on realistic stage probabilities. Total amount uses the full value. Either way, stale close dates distort both.

This Monday, pull the deals in the current quarter with a close date in the past and send each rep their list. If the list is long, you already know why the forecast moves in the last week, and you have a quarter to change it.