Calibrating HubSpot deal stage probabilities from your own win rates

HubSpot’s default sales pipeline starts with five numbers under its open deal stages: 20, 40, 60, 80 and 90 percent, per HubSpot’s Knowledge Base article Set up and manage object pipelines (last updated October 22, 2026). Your weighted pipeline multiplies each open deal by its stage’s number. Until someone replaces those numbers with stage-to-close rates from your own closed deals, the weighted figure on your board is a guess carried to two decimal places.

Calibrating them is a short job. For each stage, take the closed deals that reached it and divide the wins by the total. Then check the one HubSpot behavior that decides whether your new numbers reach the forecast at all: a deal whose probability anyone edited by hand stops following its stage. (If you want the wider picture first, why HubSpot forecasts miss introduces the default-probability problem alongside the other causes.)

Where do HubSpot’s default deal stage probabilities come from?

HubSpot’s default sales pipeline has seven stages with preset probabilities: Appointment scheduled at 20%, Qualified to buy at 40%, Presentation scheduled at 60%, Decision maker bought-in at 80%, Contract sent at 90%, Closed won at 100% and Closed lost at 0%, according to HubSpot’s Set up and manage object pipelines article, last updated October 22, 2026. HubSpot presents them as defaults. Nothing on the page says they come from your deals.

Line up the five open stages from that same HubSpot article (October 22, 2026): Contract sent 90%, Decision maker bought-in 80%, Presentation scheduled 60%, Qualified to buy 40% and Appointment scheduled 20%. New HubSpot portals start with the same 20% to 90% ladder, whatever their sales motion.

A team closing quick transactional deals and a team working long enterprise cycles get identical numbers on day one.

What the probability changes in HubSpot

The stage probability is the multiplier behind HubSpot’s weighted numbers. In board view, HubSpot calculates the weighted amount by multiplying the total amount in each stage by that stage’s probability. In the forecast tool, the Forecast deal amount setting defaults to Weighted amount, the Amount multiplied by the Deal probability, for existing and new users with Forecast permissions. Change the stage probability and both numbers move with it.

Those details come from HubSpot’s Set up and manage object pipelines article (October 22, 2026) and its Set up the forecast tool article (last updated September 9, 2026). The forecast setting’s other option is Total amount, the full deal amount.

HubSpot also has a separate property called Forecast probability, a custom percent probability that the deal will close. Calculating it automatically takes a workflow with the Set a property value action, which is limited to Professional and Enterprise accounts. For expected revenue on an individual deal record, HubSpot recommends Weighted amount over Forecast amount (HubSpot’s default deal properties, last updated September 29, 2026).

Put those together and the stage probability sits underneath your board’s weighted totals and, unless someone changed the setting, your forecast tool’s numbers too. If your forecast calls lean on categories like commit and best case, the forecast categories glossary defines them, and how to measure sales forecast accuracy covers checking whether your forecast holds up.

How do you calculate stage probabilities from your own win rates?

For each open stage, take the deals that entered it and have since closed. The share that closed won is that stage’s probability. HubSpot records Date entered and Date exited properties for each stage, which need a Professional or Enterprise subscription (HubSpot’s default deal properties, September 29, 2026). An “All” funnel report also shows conversion rates between stages (Create custom (legacy) funnel reports, November 19, 2025).

The count works in five steps:

1. Pick a window of closed deals. Two to four recent quarters is a sensible start: long enough to give each stage a real count, recent enough to reflect how your team sells now. Use closed deals only. An open deal hasn’t told you anything yet.
2. Count the closed deals that entered each stage. Filter your closed deals on the Date entered property for that stage. Won and lost both go in the count.
3. Count how many of those closed won. Same filter, plus a closed won outcome.
4. Divide wins by the total for each stage. That rate is the stage’s calibrated probability. Round it to a clean number so nobody mistakes it for false precision.
5. Enter the numbers in pipeline settings. HubSpot’s Deal probability dropdown accepts custom values: enter the value, click Add option, then select it. Keep both a Won and a Lost stage. HubSpot says you need both for sales reports, custom deal or revenue reports and sales analytics tools to process deals correctly (Set up and manage object pipelines, October 22, 2026).

Then put each of your results next to the default for that stage. The gap between them is how far your weighted pipeline has been off at that stage, in that direction.

When to trust the number. If a stage has only a handful of closed deals behind it, widen the window before you enter its rate. If two neighboring stages come out within a few points of each other, they’re probably measuring the same moment in the deal. That’s a conversation about stage definitions, and it’s worth having before you set either probability.

Why didn’t my weighted pipeline change after I updated the probabilities?

Some of your open deals stopped following their stage. HubSpot’s default deal properties article (September 29, 2026) says that once a deal’s probability is edited manually, whether on the record, by bulk edit, import, a data sync integration or the API, it no longer updates from the pipeline’s open stages. HubSpot sets it automatically again only when the deal moves to Closed Won or Closed Lost.

Recalibration skips those deals. You change Contract sent in pipeline settings, and a deal whose rep once typed in their own number keeps that number until it closes.

To find them, build a deal view for one stage at a time, add the Deal probability column, and look for deals whose probability doesn’t match the number set for that stage. Then ask where the edits came from. A rep overriding one deal is a coaching conversation. An integration or import writing Deal probability across hundreds of deals is a setup question, and you want the answer before you touch the values, or the next sync writes over your work.

Watch one detail when you correct them. By HubSpot’s own definition, setting a value on the record or by bulk edit counts as a manual update, so a deal you correct by hand still won’t move with its next stage change. Correct the values, then recheck those deals at each pipeline review until they close. Deals that have stalled in a late stage carry that stage’s probability into the weighted total the whole time they sit there, which is why stalled deals in HubSpot deserve their own pass.

How often should you recalibrate?

Recalibrate on a fixed cadence, such as each quarter, and any time you change who you sell to or how you define stages. For context, Clari Labs’ January 2026 research, a Censuswide survey of 400 North American enterprise revenue and IT leaders published by Salesloft, found 39% of organizations recalibrate forecast models only weekly or monthly. That figure is about forecast models, not stage probabilities, and it comes from vendors that sell forecasting.

Forecast accuracy has been a sore spot for years. In InsightSquared’s May 2021 press release on its State of Sales Forecasting research, CEO Todd Abbott said only nine percent of respondents were achieving a forecast accuracy of five percent or better. That’s 2021 data from a forecasting vendor, so read it as a long-running pattern rather than a current benchmark. Stage probabilities are one input to that accuracy you control directly.

A simple recalibration rule:

• Every quarter, rerun the count on the most recent closed deals.
• When you add a segment or change pricing, because deal sizes and cycle lengths move with both.
• When you add, remove or redefine a stage, because the old rates describe a pipeline that no longer exists.
• When a stage’s actual rate drifts well away from its setting at the quarterly check, even if nothing else changed.

The count gets harder when the stage history underneath it is unreliable: deals that skip stages, or integrations quietly writing probabilities. AeolusGTM is CRM, RevOps and GTM advisory, with Atlas connecting to your CRM to find gaps the CRM can’t show natively. If you want a first look at how your portal is set up, start with the HubSpot audit tool.

This week, export the last two quarters of closed deals, count how many that reached Contract sent closed won, and put that rate next to the 90% HubSpot sets by default (Set up and manage object pipelines, October 22, 2026). If the two numbers are far apart, you know which stage to start with.