Google automateddiscounts
Week one of Google-set prices: roughly break-even in aggregate — hiding big per-product winners and losers.
WooCommerce · Nordic baby gear · July 2026 · Neutral overall, wild per product
Week one of Google-set prices: roughly break-even in aggregate — hiding big per-product winners and losers.
WooCommerce · Nordic baby gear · July 2026 · Neutral overall, wild per product
Letting Google adjust prices within set floors lifts profit — conversion gains outweighing the markdowns.
Enrolled the catalogue in Google automated discounts with per-product price floors.
≈ break-even
Aggregate week-one result: approximately break-even. Underneath, product-level outcomes were wildly uneven — some products sold meaningfully more at modest markdowns, others gave margin away for nothing.
Judge the program per product, never in aggregate — break-even was the average of good and bad, not the truth about either. The floors are the real control surface. And the program is only judgeable if your margin data is true: bidding and pricing tools inherit every error in your cost feed.
One week, program-level rollout without a holdout group. Aggregate profit math depends on cost data that we separately found to be wrong in the ad-bidding feed — the same trap applies here.
Order-level revenue and margin joined to discount events, split per product; program aggregate compared against per-product distribution.
This write-up lives in the knowledge base Ankka maintains for the store it ran on — next to the queries that proved it. The free tracking audit is where your store's library starts.