Comparing every SKU every day is expensive, so routine checks can sample instead: pick a few representative SKUs daily and compare price and availability across the product page, feed, cart and checkout. The sample stays small but fixed, which is what makes a systemic feed-not-re-run drift visible.
Record results as they are; a day with no movement is not a reason to change the definition or add spend. When a mismatch appears, first rule out cache, timing or source-of-truth issues before escalating. Keep the date and source with each check so the introducing step can be traced.
This article makes no price-compliance judgement and compares no absolute amounts across shops; it only describes the local fact of whether one SKU agrees across touchpoints.