Gateway SKUs: how to test whether a product really brings in the basket
Gateway is one of the five jobs in Every SKU has a job™. It is the easiest to claim and the hardest to prove. This guide is about the proof.
By Chris Avery · Updated 28 September 2026
Advertised versus sold
With conversions with basket data, Google Ads can show the product that was advertised and the products that were sold in the same order. That lets you see when a click on one product ends in a basket of others. It does not mean the clicked product gets no credit; read the advertised and sold views side by side.
Sources: Google Ads Help, About conversions with basket data and Metrics available with conversions with basket data.
Measured, unknown, decision
- Measured
- What was advertised, what sold, basket contribution and ad cost.
- Unknown
- Whether the basket would have happened anyway, and repeat behaviour.
- Decision
- Keep spend within current economics until a test settles the unknown.
A stronger commercial case
A product can be associated with a larger basket without causing that basket. Before making customer acquisition part of its commercial role, the team needs measured repeat or basket economics and a clear view of contribution. A before-and-after comparison cannot separate advertising from seasonality, promotions or changes in demand; causal claims need an appropriate controlled test.
We help clients distinguish what the account reports from what the business can reasonably conclude, so the conversation about spend is grounded in evidence rather than assumed lifetime value.
Next
Read the BOI® overview for the five commercial jobs, and see the BOI® guide for the wider method.
Questions
- What is a Gateway SKU?
- A product whose job is to bring in new customers or support basket value, where that effect has been measured. Gateway is one of five BOI® jobs, alongside Scale, Profit, Protect and Recovery.
- Does the clicked product get no conversion credit?
- Not necessarily. With basket (cart) data, Google Ads can report which products were advertised and which were sold, so you can see when a click on one product led to a sale of others. Read both views before concluding anything.
- Does a big basket after an accessory click prove the accessory caused it?
- No. The association suggests an investigation. It does not show the basket would not have happened anyway. Incrementality needs a controlled test.
- Can I bid higher on a Gateway product because customers might come back?
- Not on an assumption. If repeat purchase is unknown, it cannot justify additional spend. The commercial case needs measured repeat or basket economics.