Kroger Household Lift Tests Designed Before Flight

More than nine in ten Kroger transactions carry a loyalty identity, which is why a matched-household control group can be built here and cannot be built at most US retailers. This engagement uses that and little else: a test designed and written down before the flight, control households matched on prior category buying, and a read that either justifies next quarter's budget or does not. If the arithmetic says your budget cannot resolve the effect you care about, you get that in the first deliverable rather than after the money is spent. The first thing that happens is arithmetic: Before anything is booked I calculate the smallest lift the test can detect given your household base, your category penetration and the flight length you have in mind. Most food and personal care brands at Kroger can resolve a low single-digit lift over an eight-week flight inside the loyalty base. A two-week burst around a promotion resolves nothing at all, whatever the platform reports afterwards, and I would rather tell you that before you commit than write it up as a learning later. What gets bought, and against which question: Onsite search and category placements on Kroger.com and in the app, offsite display and video, and in-store surfaces each answer a different question, and they do not go into one test cell. One question per quarter: whether onsite defends your basket against a private label push, whether offsite buys households new to the brand at a price you would pay again, whether an in-store surface does anything measurable at all. Cells stay fixed once the flight starts. Moving budget between cells mid-flight ends the test, and I will say so out loud rather than quietly patch the analysis. The read: Test against matched control at household level, with the pre-registered design attached so nobody relitigates the definitions afterwards. You get incremental units and incremental spend per household, cost per household new to the brand, repeat rate among exposed households against control at eight and sixteen weeks, and a written recommendation on next quarter's number. Where 84.51 reporting and your own shipment data disagree, the disagreement goes in the document rather than being resolved in favour of whichever number is friendlier. What I will not accept: A lift study designed after the flight has run. A test window that also contains a national television flight or a chain-wide feature and display event, unless the control group sits inside the same conditions. A result read on the promoted item alone when the whole category was on deal. And percentage-of-spend compensation, which would make me the wrong person to tell you the answer is to spend less at Kroger next quarter. Not included: Creative and asset production, shopper marketing execution, promotion funding, category management deliverables for your buyer, and day-to-day campaign operations where you already have an agency in the account. In that case they fly it and I design and read it, which works well and keeps me clear of judging work I booked myself. Turn me down if: Your Kroger business runs under about a hundred thousand dollars a quarter, where a household test cannot carry the cost of designing it properly. You need a decision inside three weeks. Or you are planning on the assumption that Kroger and Albertsons will combine: that merger was blocked in December 2024, and the two networks should be planned and measured as separate buys.