Weighted pipeline forecasting: a sales forecast a small team can actually trust

A weighted pipeline forecast estimates revenue by multiplying each open deal's value by its probability of closing, then adding the results. A $20,000 deal at 25% counts as $5,000; a $4,000 deal at 75% counts as $3,000. Sum every deal expected to close in the period and you have the forecast.

The method is simple and widely used because it turns a pipeline full of uncertain deals into one number. Its accuracy depends entirely on two inputs: probabilities that reflect how your deals really close, and close dates that reflect when. This guide covers how to set both, how to read the result, and how a small team keeps the forecast honest from one month to the next.

The arithmetic, with a worked example

Take a pipeline of seven deals expected to close this quarter. Stage probabilities are set from the team's history (more on that below).

The deals and numbers are invented for illustration. The pipeline is worth $78,900 at face value; the weighted forecast for the quarter is about $30,000. That gap is normal. Planning hiring or spending against the face value is a frequent and expensive mistake.

Where the probabilities should come from

Most CRMs ship with stage probabilities such as 10%, 25%, 50%, 75% and 90%. Those are placeholders. A forecast built on them is a forecast built on someone else's sales process.

The better source is your own history. For each stage, look at every deal that reached that stage in the last two to four quarters and has since closed, and calculate the share that was won:

stage probability = deals won after reaching the stage ÷ deals closed (won + lost) after reaching the stage

Three practical notes make this work:

• Use closed deals only. Deals still open have not resolved yet and would pull the rate down.
• Count by stage reached, not stage lost at. A deal won from Negotiation also passed through Proposal and counts towards Proposal's rate.
• Accept rough numbers with small samples. With thirty closed deals a quarter, rates will move a few points each time you recalculate. Round to the nearest 5% and revisit quarterly.

This only works if outcomes are recorded separately from stages, so you can see where a lost deal actually stopped. If Lost is a column on your board, fix that first: sales pipeline stages with exit criteria explains why.

Stage probability or deal probability?

There are two ways to assign a probability to a deal, and teams often mix them without noticing.

• Stage probability: every deal in a stage gets the same number. It is consistent, cannot be gamed by an optimistic owner, and improves as history accumulates.
• Deal probability: the owner sets a number per deal. It captures what the stage cannot know (a champion left, a competitor dropped out) and is exposed to optimism.

A workable compromise is to default each deal's probability from its stage and let owners override it with a reason. Review overrides in the forecast meeting: a team whose overrides are consistently higher than the stage rate is telling you something about either the stages or the team.

Reading the forecast without fooling yourself

A weighted forecast is an expected value. It is accurate across many deals and noisy across few. Four checks keep it useful:

1. Look at concentration. If one deal makes up 40% of the weighted total, the forecast is really a bet on that deal. Review the largest deals one by one and ask what would have to be true for each to close this period.
2. Separate commit from weighted. Ask owners for a "commit" figure, the deals they would stake their name on. The weighted number is the statistical view; commit is the human one. A large gap between them is a conversation worth having.
3. Watch stalled deals. A deal that has sat in Proposal for three times the usual duration does not have Proposal's probability any more. Either move it forward with a concrete step or lower the probability.
4. Check close-date slippage. Count how many deals had their close date pushed to the next period. A forecast that keeps rolling the same deals forward is overstating the current period every time.

Calibrate against what actually closed

A forecast becomes trustworthy when you check it. At the end of each month or quarter, write down three numbers: the weighted forecast at the start of the period, the commit figure, and the revenue that actually closed. After three or four periods a pattern appears.

The second row matters for teams with short sales cycles. If many deals are created and won inside the same month, the pipeline at the start of the month will always underestimate the month. Adding the average of recent "created and closed in period" revenue as a separate line fixes that.

When to use a different forecasting method

The weighted pipeline is the right default for B2B teams with distinct deals and a sales cycle of weeks or months. It is not the only method, and in some businesses it is the wrong one.

• Run-rate forecasting projects recent monthly revenue forward, adjusted for seasonality. It suits high-volume, short-cycle sales, such as e-commerce or small subscriptions, where individual deals are too small and fast to track in a pipeline.
• Category forecasting asks owners to place each deal in commit, best case or pipeline. It is quick and captures judgement, but it is only as good as each owner's honesty. It works well alongside the weighted figure, as described above.
• Cycle-length forecasting uses how long deals usually take from creation to close to estimate when today's early-stage deals will land. It is useful for planning beyond the current quarter, where close dates are guesses.
• Recurring revenue from existing contracts should not be in the new-business pipeline at all. Forecast renewals and expansions separately, with their own probabilities, because they behave very differently from new deals.

Most small companies end up with the weighted pipeline for new business, a run-rate line for small fast deals, and a renewals list. Three simple numbers added together are easier to explain and correct than one sophisticated model.

How Nextora shows the forecast

In Nextora a deal record holds the four values a forecast needs: stage, amount, probability and close date. The pipeline board works out three figures at the top: money in flight, won this month and the probability-weighted forecast, plus how the money is spread across stages. Each column's footer shows the stage total and its forecast, so the weighted view is visible stage by stage, not only as one figure.

Outcome is a separate field from stage, so a lost deal stays at the stage where it stopped and the history needed for stage win rates is there, while deals on hold are excluded from conversion. Pipeline stages with exit criteria explains that model.

Pipeline analytics, computed from the history of stage moves, shows the win rate for a period and a Time in stage table with each stage's median and the count of deals past their threshold: the stalled deals from check 3 above. Analytics also offers a ready-made "Sales: pipeline and money" dashboard calculated from your own data. The analytics page shows the dashboards, and the CRM page the board.

A monthly forecasting routine

1. First working day: record the weighted forecast and the commit figure for the month.
2. Every week: fix missing and past close dates, review the five largest deals, and clear the stalled list.
3. Last working day: record what actually closed and the deals that slipped.
4. Every quarter: recalculate stage probabilities from closed deals and compare forecast to actual across the quarter.

A forecast rests on clean deals, and clean deals rest on a clear record model; the five records a CRM actually needs is the foundation. When won deals turn into orders and invoices, the next number to forecast is cash: quote to cash for a small business picks up from there.