Introduction
Open almost any CRM and there is a weighted pipeline number waiting for you. Discovery counts for 20 percent, proposal for 60, negotiation for 80, and the system quietly sums it all into a figure that looks like a forecast. It gets read out on calls. It goes into board decks.
AI tools promise to replace that figure with a prediction, and we wrote separately about AI forecasting for a small team and what it needs from your data.
It is not a forecast. It is a description of the shape of your pipeline, dressed in the language of prediction. This article explains what weighted forecasting actually computes, the three structural reasons stage percentages mislead, what to use instead, and where weighting still earns its keep. For what a CRM can and cannot tell a manager about next quarter, see revenue forecasting for a sales manager.
What a Weighted Pipeline Forecast Actually Calculates
The arithmetic is simple. Every deal is multiplied by the probability assigned to its current stage, and the results are added up. A $100,000 deal in a stage weighted at 60 percent contributes $60,000. In Pipedrive there is a wrinkle: a deal’s own probability overrides the stage’s, so the multiplier is not always the number on the stage.
Read that literally and you can see what it means: the number is the expected value of the pipeline if every deal in a given stage behaved like the historical average of that stage. It is an average of averages. For a portfolio of many similar deals that can be a reasonable summary. For the forecast of a specific quarter, in which a handful of large deals decide the outcome, an average of averages is the wrong tool — it describes the middle of a distribution while you are being asked about one particular draw from it.
There is also a subtler issue. No single deal will ever close for 60 percent of its value. Each one lands at 100 percent or zero. The weighted number is a blend of outcomes that cannot occur, which is fine as a summary statistic and misleading as a commitment.
Why Stage Percentages Lie: Three Structural Problems
1. The percentages average deals that behave nothing alike
A single number per stage assumes every deal in that stage has the same chance of closing. In practice, two deals sitting in proposal can be entirely different animals: one has a confirmed budget, an engaged sponsor and a signed-off next step; the other has been re-forecast twice and has not had a reply in three weeks. The model gives them the same probability.
Deal size makes this sharper. Large deals typically convert at a different rate and on a different timeline than small ones, so applying the same percentage to a $250,000 opportunity and a $15,000 one bakes in an error that grows with the number that matters most.
A single large deal at 60 percent still distorts the total; the fix is to show a range, as we do with one big deal in the forecast.
2. The percentages are set once and never recalibrated
Ask where your stage percentages came from. In most teams the honest answer is that someone chose them during implementation, possibly from a template, and nobody has revisited them since.
Meanwhile the business changed. You moved upmarket, or added a security review, or changed the definition of what qualifies for proposal. Every one of those changes moves the true conversion rate of a stage, and none of them updates the number in the CRM. A model calibrated to a business you no longer run will be confidently wrong.
3. Weighting rewards volume of early-stage pipeline
This is the failure mode with real behavioural consequences. If early stages carry a non-zero weight, every deal added to the top of the funnel raises the weighted forecast, regardless of quality.
That is the difference between a funnel and a pipeline showing up in the forecast.
A team that is behind can improve its forecast by creating opportunities. No customer conversation has changed, nothing is more likely to close, but the number goes up — and the incentive to keep marginal deals open rather than close them lost points the same way. The forecast becomes something that responds to the pipeline’s size rather than its health, which is precisely backwards.
Better Alternatives to Stage-Percentage Weighting
None of these are harder than what you are doing now.
Commit and best case categorisation
Ask the rep to place each deal into a category rather than inherit a stage number: commit for deals they will stake their name on, best case for deals that need something to go right, pipeline for everything else. The value is that judgement becomes visible and personal, and categories can be reviewed against outcomes afterwards. Over a few quarters you learn what each rep’s commit is actually worth, which is far more useful than a stage average.
Our guide to forecasting with commit and best case instead of stage percentages covers the entry rules for each category.
Historical stage conversion measured for your own team
If you want percentages, derive them from your own closed deals rather than accepting the defaults. Take the last several quarters, measure what share of deals that reached each stage eventually closed, and segment by the dimension that matters most for you — usually deal size or product line. Recalculate every quarter or two. This is the same model, honestly calibrated, and it fixes the second problem while leaving the first partly intact.
Whether that is worth doing depends on how many closed deals you have, and we map the choice in sales forecasting methods by closed-deal count.
Judgement backed by inspection
The most accurate forecasts in most teams come from a manager who reviews the specific deals that decide the number and asks the same questions each time: is there a confirmed next step, has the economic buyer been involved, what has moved in the last two weeks, what is the compelling event. This does not scale to every deal, but it does not need to. Twenty deals usually decide the quarter.
Worked Example: One Pipeline, Three Forecasts
Consider a quarter with a $500,000 quota and a pipeline of ten deals. Figures are illustrative.
Three deals worth $80,000 each sit in negotiation, weighted at 80 percent. Three worth $100,000 sit in proposal, weighted at 60. Four worth $70,000 sit in discovery, weighted at 20.
Method one, stage weighting. Negotiation contributes $192,000, proposal $180,000, discovery $56,000. Weighted forecast: $428,000. Comfortably short of quota but within striking distance, and the shape of a quarter nobody panics about.
Method two, categories. The rep commits two of the three negotiation deals and one proposal deal, because the third negotiation deal lost its sponsor last month and two of the proposals have no confirmed next step. Commit is $260,000, best case adds $180,000. The picture is now a $260,000 floor with real work to do — a materially different quarter from the one method one described.
Method three, own conversion rates. Measuring actual history, negotiation converts at 65 percent rather than 80, proposal at 35 rather than 60, and discovery at 8 rather than 20. The same pipeline yields $156,000 plus $105,000 plus $22,400, or about $283,000.
We wrote separately about calculating win rate from your own closed deals.
Three numbers from one pipeline: $428,000, $260,000, $283,000. The two honest methods agree with each other and disagree with the CRM default by roughly $150,000 — about a third of the quota. That gap is not noise. It is the cost of never recalibrating.
Weighted Forecasting Is Not Useless — It’s Just Not a Forecast
Weighting is a fine descriptive statistic. It is a reasonable way to compare this quarter’s pipeline shape to last quarter’s, to see coverage moving across stages over time, or to give a rough read on a large portfolio of similar transactions.
What it should not do is decide the number you commit to. Use it as a dashboard summary; use categories, your own conversion history, and inspection of the deals that matter for the number you say out loud. Read the columns of the forecast view the same way, and keep your hands off the cards while you do.
For a short list of which numbers to keep on the dashboard next to it, see the manager’s dashboard guide.
Whichever method you choose, the weakness is the same: the pipeline going into it. Stale deals, missing next steps and close dates that have already passed corrupt every forecasting method equally. This is where continuous monitoring helps — Pipeline Guard gives each active deal a live health score and flags stalled and silent deals, missing next steps and overdue activities, so the pipeline behind the forecast is worth forecasting from in the first place.
FAQ
What is a weighted pipeline forecast?
The sum of every open deal’s value multiplied by a probability assigned to its current stage. It answers what the pipeline would be worth if each deal behaved like the average of its stage, which is not the same as what will close this quarter. Note that the weighted number is only as good as the amount in Value, and with three package prices that amount needs a rule. With deals in more than one currency, it also inherits the limits of mixed-currency totals.
Are stage percentages ever accurate?
They are accurate as averages over many similar deals when derived from your own recent history and recalibrated regularly. They are least accurate exactly where you need them most: a specific quarter decided by a few large deals.
How often should stage conversion rates be recalculated?
Quarterly is a reasonable rhythm, and immediately after any change to stage definitions, qualification criteria or target segment — all three invalidate the old numbers.
Should early-stage deals carry any weight at all?
For in-period forecasting, usually not. If a deal cannot realistically close in the period, including a fraction of it inflates the number and rewards pipeline creation over pipeline progression.
A rep who keeps such deals in the period anyway is showing one of the two ways a rep’s forecast is wrong.
What single change improves forecast accuracy most?
Requiring a confirmed next step with a date on every open deal. It costs one field and it removes the largest category of deals that are forecast as alive while being effectively dead.
Conclusion
Stage-percentage weighting does not lie because the arithmetic is wrong. It misleads because it answers a question about averages while being read as an answer about your quarter, using percentages that were probably never calibrated to your business and that quietly reward adding pipeline over advancing it.
Keep the weighted number as a description of pipeline shape. For the number you commit to, use categories, conversion rates measured on your own closed deals, and inspection of the deals that decide the quarter. Start this week by calculating the real conversion rate of one stage from your own history and comparing it to the percentage in your CRM — the gap is usually the whole story.
To see how fast that pipeline turns into revenue, track sales velocity alongside it.
Related reading
- Sales Forecast Accuracy: How to Measure It and How to Raise It
- Pipeline Coverage Ratio: How to Calculate It and What to Trust
