Control group
Updated September 30, 2026
In one sentence
A control group is a few customers, drawn at random, who don’t receive a promotion. By comparing them with the others, you see what the promotion really added.
Why keep one
Without it, it’s easy to get fooled. Some of the customers who come back after a message would have come back anyway. Compare those who got the offer with those who didn’t, over the same period. Then you see the effect of the offer alone, without the season, the weather or a holiday.
How to set it up
Draw customers at random, never hand-pick them. A small share is enough. For example, the email tool Klaviyo suggests about 5% of contacts for a group left out of all sends. With few customers, the gap from a single campaign can be chance. Look at the trend across several campaigns instead.
In the card
Every campaign can keep a control group. The report then compares both groups’ visits in the days after the send. It shows you the visits gained from the offer, not all visits.
A restaurant sends an offer to 900 members and randomly keeps 100 without a message. Within 7 days, 12% of the members who got the offer come back, versus 8% of the control group. The offer added about 4 points, or 36 extra visits for 900 sends.
See your business’s card
Send us your logo: we prepare a preview of your card and show it to you on video.