Platforms
Your platform decides how hard the data is to get, not what the decision should be.
Shopify, WooCommerce and Adobe Commerce each expose product, margin and stock data differently. The commercial question is identical on all three: what is this SKU's job this week, and does the data support that answer.
Why this changes the advertising decision
What is different here isn't the channel. It's the economics underneath it.
Shopify: flexible catalogue, thin cost data
Variants and collections are easy to change and easy to lose track of. Cost and margin usually live outside the platform, so contribution has to be reconciled before it can drive a bid.
WooCommerce: variable products, inconsistent feeds
Parent and variation data often disagree. Bidding at parent level buries the variants that are actually losing money.
Adobe Commerce / Magento: large catalogues, deep attributes
The data is there and the volume hides the problem. With thousands of SKUs, loss-makers survive for months inside a healthy blended number.
All three: availability changes faster than the feed
Stale availability is the most common cause of paid clicks landing on something that cannot be bought.
What we do about it
The execution follows the economics, not the other way round.
Contribution data is reconciled, not assumed
Cost, shipping, returns and fees are brought together into a contribution figure per SKU, wherever they actually live.
Bidding happens at the level demand happens
Variant-level demand is bid at variant level rather than averaged up to the parent product.
Availability is kept current
Feed refresh cadence is matched to how fast your stock actually moves, so spend does not chase unbuyable SKUs.
The platform is an implementation detail
The commercial logic is the same on all three. Only the engineering to get the data out differs.
The Method - BOI®
How BOI® applies here.
The problem
Platform choice changes how hard the data is to reach, not what the allocation decision should be.
The BOI® answer
Whatever the platform, each SKU carries one job: Scale, Profit, Protect, Recovery or Gateway. The feed's only task is to make that job legible to Google.
Catalogue definitions
Products, variants, feed offers and active SKUs are different numbers
A “large catalogue” can mean any of these. Before comparing accounts, or reading a SKU count, check which unit is meant.
- Product
- A distinct thing you sell, such as one style of shirt, regardless of size or colour.
- Variant
- A sellable version of a product, for example the shirt in medium, blue. Often the level stock is held at.
- Feed offer
- One item submitted to Merchant Center. Depending on the feed, that may be a variant, a product or a regional or local offer, so the count can exceed products many times over.
- Active advertised SKU
- An approved offer that is eligible and actually receiving budget in a period. Usually far smaller than the feed, and the number that matters for management workload.
Large-catalogue work in our published cases
- UKSoccerShop
The award entry title refers to 6.43M SKUs; the case describes custom labels segmenting 2,000+ SKUs. These are different units and should not be compared or summed.
- Closure London
A Shopping routing example covering 591 SKUs across four product roles, within the wider account result.
- Multi-brand retailer
Catalogue segmented into five commercial tiers. The client withheld its name; the SKU count is not published.
Our published cases do not name the client's ecommerce platform, so we do not present any of them as platform-specific evidence. Results in each case are observed outcomes with the stated source and caveats, not guaranteed or causal lift.
Next step
We'll tell you what your account is doing to your ecommerce platforms margin.
A commercial review reads your account against your own product economics, and tells you where the next pound should go. Brands we work with typically spend around £10k a month on Google Ads, but that is guidance, not a gate. If you are below it and the question is real, ask anyway.
Related industries