Win Rate = Won over (won + lost) is Not What You Think it is

Won over (won + lost), aka the "WOWPL" (won over won plus lost) formula: a bad estimate of something you don't want.

This is pretty geeky, but win rates are the foundation of sales planning, forecasting, and optimization. Let's get to it.

Last week I posted that 80% of businesses are using the flawed WOWPL formula.

Someone asked me yesterday "What's wrong with WOWPL? Why is it flawed."

It's simple. WOWPL relies on accurate recording of wins AND losses.

• You know for certain when you win. Wins are reliably recorded when they happen.
• You don't know for certain when you lose. Losses are cleaned up episodically. Or maybe never.

Uncertainty about when losses occur makes for big variations in the WOWPL formula... particularly if you are measuring over short periods, or periods with little data.

If you measure over periods with a lot of data and long enough to capture a consistent amount of loss cleanup in each measurement period, then WOWPL can be a reasonable but optimistic (because it omits open deals) estimate of your long-term win rate. That is, if you wait for ever.

• What's enough data? Easy, do statistical tests of significance when you compare two measurements. Do you do that?
• What's long enough? How optimistic? Great questions! Depends on your business practices. How long do you measure over?

What should you use instead? Use the time to event method. TTE is completely insensitive to these problems.

TTE only cares about when a deal was created and when it was won (if it was won). Timing of recording losses don't matter. So TTE is much?less affected by bad data hygiene. Estimates can be tuned to be?responsive?to short-term win-rate trends. The time-to-event approach is also?insensitive?to changes (even dramatic changes) in deal generation rate. Here's an explainer blog with an Excel workbook TTE example.