Proving a campaign actually worked

Do personalised incentives actually bring dormant customers back, or do those customers come back anyway? And if the effect is real, is it worth scaling? The business needed the incremental impact — not just a count of how many lapsed customers happened to return.

The client is a high-volume on-demand delivery platform with a meaningful population of dormant customers — people who had ordered before but gone quiet. Some lapsed customers always return on their own.

That makes it dangerously easy to pay for reactivations that would have happened regardless, and to mistake natural return for campaign success. Nobody had measured whether the spend was actually causing anything.

Dormant customers were grouped by how recently they had lapsed, and within each cohort split into a test group that received a free-delivery incentive and a matched control group that received nothing. The incentive went out through a single channel — push notification — to keep the comparison clean, with daily list refreshes and already-reactivated customers excluded.

Because every test customer had a statistical twin in the control group, any difference in reactivation between the two was incremental: caused by the campaign, not by customers returning on their own.

The campaign worked, and we could isolate the part that was genuinely incremental. The test group reactivated at a meaningfully higher rate than the control and generated about a third more orders. The effect was strongest among the most recently lapsed customers — the people most worth targeting.

Critically, around 90% of the incentive cost was recovered within the first two weeks, which made the economics defensible rather than hopeful.

For the first time, the business knew the campaign drove real, incremental reactivation, knew roughly what that uplift was worth, and knew which dormant customers responded best — so any scale-up could be aimed at the cohorts where the return was highest.

Just as importantly, it now had a repeatable experimental method for testing the next campaign before committing real budget to it.

Marketing Experimentation is one of five Hira Deep-Dive Sprints — a focused, four-week engagement that measures whether marketing spend actually causes the outcome it claims, using proper test-and-control design. It's the right fit when you need to know what a campaign really moved before you scale it.