The Flowup Insights Team put this how-to together for people who are about to approve a 7-Eleven campaign report that only shows a before-and-after line. It is a practitioner note from the Flowup team, not a copy of a web guide.
Convenience store sales move with holidays, weather, seasons and what sits next to you on the shelf. A busy month will make almost any activity look effective, even when the same rise shows up in stores that saw none of your work. The usable method is to compare how sales per store per day changed in an area that ran the work, against a similar area that did not.
Match the two areas before anyone spends
Pick areas with a similar number of stores that actually have stock, and similar sales per store per day before the test. Keep store types alike: office districts with office districts, not with residential streets. Avoid a pair where many new stores will open during the test.
TikTok ads can be aimed by location, so they can respect the split. Creator videos travel more widely; treat them as a shared effect, not as a treatment that stayed inside the test area.
Pull at least several weeks of history, ideally four to eight, and check that the two areas already moved together. If they did not, change the pair before you spend.
Change only one thing, then do the subtraction
During the test, the only planned difference should be campaign or no campaign. If price, an in-store promotion, placement or packaging also changes, you will not know what moved the sales. If a change cannot be avoided, make it in both areas and write it down.
Then look at the change in each area, and subtract. A hypothetical example, only to show the arithmetic: the test area goes from 0.40 to 0.52 units per store per day (plus 0.12); the control area goes from 0.42 to 0.46 (plus 0.04). The difference you can reasonably credit is about 0.08, not 0.12. Using only "before versus after" would overstate the effect by half. Do not treat these figures as a target.
Write down disturbances as they happen: stock-outs, a competitor's promotion, a long holiday, news in the category. They do not automatically void the test, but they explain odd weeks.
The pairing rules, the worked subtraction, and what to do when you cannot get data by area are on Measuring 7-Eleven marketing with test and control areas. Open that page when you name the two areas and build the one-line result for the meeting.
If the report you were sent has no control area, do not sign off on "it worked." Ask for a pair, or say clearly that you are looking at a before-and-after only.
Flowup Insights Team ยท a how-to for reading 7-Eleven campaign results without fooling yourself
Top comments (0)