Why does Datapeople exclude active candidates from the Drop-off calculation?

There are two main reasons why we exclude active candidates from our drop-off calculations:

1. As candidates get resolved, the drop-off rate would change
2. The true drop-off rate can be masked (made to look smaller) by comparing to a large group of candidates, many of whom will eventually advance to other stages in the process or be resolved with a rejection/hire/etc.

We analyzed millions of recruiting outcomes to develop the Net Drop-off Rate. Our results showed that the conventional Drop-off Rate calculation would vary as the status of active candidates are resolved. This makes the conventional Drop-off Rate a poor signal to identify issues early in processes since, for most of its life, the rate will be underestimated.

In contrast, the Net Drop-off Rate is less prone to fluctuation when you are looking at more recent recruiting processes. Active candidates are excluded because we do not know which end-state they will end up in. Rather than report misleading metrics that may hide the true story, we choose to report on the smaller, but cleaner data set that can give us the most accurate report with the data we currently have.

When we include active candidates in the dropoff rate calculation, these four companies seem to have similar dropoffs rates (orange bars). However, when we remove the active candidates from the calculation (purple bars), Companies C & D have significantly higher dropoff rates. Because they had so many more unresolved candidates than Companies A & B, including active candidates in the dropoff rate calculation masked their higher dropoff rates.