Response rate calculation | Experience Community
Skip to main content
Solved

Response rate calculation

  • June 12, 2026
  • 1 reply
  • 9 views
A
Level 2 ●●
Forum|alt.badge.img+1

I have a custom metric called Response Rate, which is calculated as: (Survey Finished) / (Email Sent - Email Bounce). These data points are available in the respondent funnel data.

Now, I need to create a new widget to display the Response Rate for shorter surveys and extended surveys. The survey type will be identified using an embedded data field (randnum) available in the survey source, where randnum > 40 indicates an extended survey and randnum < 41 indicates a short survey.

However, I am unsure how to calculate the correct Response Rate. Should I apply a filter based on randnum, or different calculation needed. 

Best answer by vgayraud

Your response rate formula stays the same, but the filter only works if randnum lives at the sample level (set on the contact before distribution), not in the survey flow.

Sent and Bounce are funnel events that happen before anyone opens the survey. If randnum is set inside the survey flow, only people who actually started get a value and every bounce and non-starter is blank. So randnum > 40 would drop all of them from your denominator and inflate the rate.

Fix: set randnum on the contact at the sample level before sending and add it to the tracked ED in thefunnel data config.
 

Then just filter the widget randnum > 40 / < 41 , same metric, no recalculation.

If randnum can only live in the flow, you can't split it correctly because non-starters never got the value.

1 reply

vgayraud
QPN Level 7 ●●●●●●●
Forum|alt.badge.img+65
  • vgayraud
  • QPN Level 7 ●●●●●●●
  • Answer
  • June 12, 2026

Your response rate formula stays the same, but the filter only works if randnum lives at the sample level (set on the contact before distribution), not in the survey flow.

Sent and Bounce are funnel events that happen before anyone opens the survey. If randnum is set inside the survey flow, only people who actually started get a value and every bounce and non-starter is blank. So randnum > 40 would drop all of them from your denominator and inflate the rate.

Fix: set randnum on the contact at the sample level before sending and add it to the tracked ED in thefunnel data config.
 

Then just filter the widget randnum > 40 / < 41 , same metric, no recalculation.

If randnum can only live in the flow, you can't split it correctly because non-starters never got the value.

Vincent Gayraud - https://www.linkedin.com/in/vincent-gayraud-790344116/
    Terms & ConditionsAccessibility statement