The same cycle-detection approach, pointed at grocery. Kroger and its banners run recurring 4x-fuel-point and gift-card multiplier events — and unlike office supply, those deals are advertised in advance.
Nothing here is live yet. The dashboard's model — observed runs recorded as data, cycle inferred from the record, confidence derived from the share of due weeks that actually ran — transfers directly. What changes is the source.
Why grocery is easier than office supply: Kroger's gift card promotions appear in the weekly circular, which is indexed by the flyer aggregators. Office supply gift card deals never are — they're in-store only and unadvertised, which is why confirming them depends on someone physically seeing the display. Grocery promos can be confirmed automatically, days ahead.
The office supply cycle is proven — four streams, 28 days, backtested across eighteen months. The grocery equivalent hasn't been measured yet, and publishing a prediction before the pattern is verified would undermine the thing that makes this site worth reading.
First job is collecting the record. The forecast comes after.