<aside> 💡 This document reviews funnel metric concepts, how to build funnels in Ashby and how to interpret the results.

</aside>

Overview Sections:

https://www.loom.com/share/12f9abbe40a3476dbc29f11f9f1305ae

Building a funnel in Ashby

Funnels are part of Ashby's standard report template library. Simply navigate to Reports > Templates and select "Funnel Report" from the library. By default, Ashby's funnel report selects all Job Considerations (for Greenhouse users: Applications, Lever users: Opportunities) with interview activity in the last 6 months.

https://s3-us-west-2.amazonaws.com/secure.notion-static.com/8ee105e8-aa38-426d-8cc3-f0c7355d38ba/Funnel_Report_(2).png

What goes into a funnel? How selections form a cohort

When Ashby's funnel report defaults to selecting all Job Considerations that had an interview in the last 6 months, a selection is taking place. That selection results in a fixed group to analyze, sometimes referred to as a cohort*.* The starting stage of the funnel shows the size of the cohort, giving a total count above the bar. In the example above, a cohort of 1,962 Job Considerations were selected for analysis in the funnel.

<aside> 💡 As a single candidate can have multiple Job Considerations it's possible that an individual candidate contributes to a funnel multiple times, and progresses to different stages per Job Consideration.

</aside>

Completed Job Considerations Only

For a funnel to be accurate, it can only analyze Job Considerations that have completed their interview process (either hired or archived). As we do not know in advance how far any consideration will progress, the funnel will be skewed if any counting occurs prematurely.

Interpreting Funnel Results

Reading our funnel report from left to right, we see that the first stage shows the total cohort size, as noted above. The subsequent stages show a percent conversion from the prior step. The 78.2% in the second stage is the percentage of the first stage that proceeded, and so on. This percentage is sometimes referred to as a "passthrough rate".