Amazon quietly rolled out a new predictive audience layer inside Sponsored Products. It is similar to what we see in Sponsored Brands.
But now it sits directly inside SP bid adjustments. It is often labeled as "High likelihood of purchase based on shopping activity." And it is likely hiding in your bid adjustment settings right now.
Here is the logic behind it.
Amazon analyzes shopper clicks, views, carts, and purchase patterns in real time.
It targets high-intent buyers instantly.
Instead of increasing bids for everyone searching a keyword, you increase bids only for shoppers who show strong purchase intent. This helps you keep your base bids lower for general traffic.
Important note about naming
Different accounts report different versions of this feature. Some see “High likelihood of purchase based on shopping activity.” Others see “High interest based on shopping history.”
There are also versions like “High purchase intent based on shopping signals” or shorter labels such as “High interest shoppers.”
Amazon is A/B testing the naming and descriptions across accounts during the rollout. All versions come from the same underlying machine learning signals and behave the same way.
The Efficiency Data
Amazon tested this across 1,700 campaigns and saw a median ROAS lift of 77% for shoppers in this audience. This does not mean you should activate it everywhere.
Strategic Application
This audience works best in broad environments where you need more precision. Think broad match keywords, auto campaigns, or category targeting.
Execution Checklist
Broad match, auto, or category targeting.
Look for the “Audiences built by Amazon” section in your campaign settings.
Select the high-intent audience and use a 20 to 30% bid adjustment. Avoid large jumps.
Compare “in audience” and “out of audience” ROAS inside your reports.
The Shift in Logic
Most brands optimize bids by placement, device, or hour. This lets you optimize based on purchase intent. It is a strong lever, but like any lever, it requires data checks to confirm it improves efficiency.
