Why Your Win Rate Keeps Changing

Three defensible formulas on the same export, twenty points apart, all called win rate. What each denominator includes, which one answers which question, and why nobody notices the switch.

Win rate is quoted in nearly every sales meeting and defined in almost none of them. There are three defensible formulas, they use the same export, and on real data they land twenty points apart. The number does not keep changing because the business is volatile. It keeps changing because different people are computing different things and calling them the same word.

The three definitions

1. Count, of closed     won deals / (won + lost) deals
2. Value, of closed     won amount / (won + lost) amount
3. Of the pipeline      won deals / all deals, open included

All three are in use in real companies. Only the first is what most people mean.

A worked quarter, one export, three answers:

Won      18 deals    1,240,000
Lost     22 deals      310,000
Open     61 deals    4,900,000

1. Count, of closed    18 / 40           = 45.0%
2. Value, of closed    1,240 / 1,550     = 80.0%
3. Of the pipeline     18 / 101          = 17.8%

Forty-five, eighty, or eighteen. Nothing is miscalculated. The team won most of the money and slightly under half of the deals, which is a completely coherent picture of a business that loses small deals and wins large ones — and depending on which figure reaches the board, it reads as a strong quarter, an exceptional one, or a crisis.

What each denominator is actually measuring

Count, of closed. How often you win when a deal reaches a decision. This is the honest read on sales execution, because it does not let one enormous deal carry a quarter. Default to it.

Value, of closed. What share of decided money you captured. Useful, and volatile — a single large loss can move it by twenty points, which makes a quarterly series jumpy in ways that do not correspond to anything anyone did.

Of the pipeline. This one is the trap. It falls whenever the pipeline grows, so a great prospecting month lowers it, and it rises whenever someone clears out stale deals. It is a pipeline-health indicator wearing win rate's name, and it moves for reasons that have nothing to do with winning.

Why it moves when nothing has happened

Four causes, none of which is performance:

• A cleanup. Two hundred stale deals marked lost in one afternoon. The denominator jumps, the rate collapses, and the sales team is asked to explain an administrative decision.
• Reclassification. Deals moved between types or pipelines. Same deals, different population.
• The date field. Bucketing by close date, created date or last-modified date gives three different quarters from one export. Close date is the right answer for win rate, and it is not always what a report was built on.
• The stage list. Whether a stage like disqualified or no decision counts as lost, as excluded, or as neither. Rarely written down, frequently changed.

Each of these makes the number move without the business moving. Together, they explain most of the mystery in most win-rate series.

Doing it in Excel

With a deal export in a table named Deals:

Won count     =COUNTIFS(Deals[Stage],"Closed Won")
Lost count    =COUNTIFS(Deals[Stage],"Closed Lost")
Win rate      =IFERROR(B1/(B1+B2),"")

Won value     =SUMIFS(Deals[Amount],Deals[Stage],"Closed Won")
Lost value    =SUMIFS(Deals[Amount],Deals[Stage],"Closed Lost")
Win rate (£)  =IFERROR(B4/(B4+B5),"")

Two traps in that snippet. The stage labels must match your export exactly — Closed Won, closedwon and Closed-Won are three different strings, and a filter that matches none of them returns a confident zero rather than an error. And the criteria must not silently exclude a stage you meant to include: total your stage counts and check they sum to the row count of the export.

For a quarterly series, add a close-date quarter column and a criterion. Use the close date, not the created date, or a deal counts in the quarter it was born rather than the quarter it was decided.

What win rate is not

It is not a lead conversion rate. Win rate has a denominator of deals that reached a decision; lead conversion has a denominator of every lead that arrived, which is very much larger and lives on a different object in your CRM. Quoting one where the audience assumes the other is off by an order of magnitude, and both parties leave the room confident.

It is also not a forecast. A rate describes what already happened to closed deals; applying it to open ones is a projection with assumptions of its own, and the honest version of that argument is in weighted pipeline is not a forecast.

One definition, held still

Quiriz for CRM defines win rate as won deals over closed deals — counts, with open deals excluded — so the same question returns the same figure whoever asks and whenever they ask. In Excel that is =QUIRIZ.ASK("win rate by sales rep this quarter", "table").

Won amount and closed amount are available as their own metrics, so the value-based view is a question you can ask deliberately rather than one you get by accident.

The honest boundary: we cannot tell you whether a deal marked lost was really lost. Stage hygiene is yours. What a stored definition fixes is the second problem — that the same word produced three numbers — and that one is worth fixing on its own.

Frequently asked questions