When most brands dip their toes into Amazon DSP, they expect a magic "revenue bu
By Michael Erickson Facchin · June 11, 2026 · Curated by George's Blog
When most brands dip their toes into Amazon DSP, they expect a magic "revenue button." But the real magic isn’t just in access to Amazon's exclusive first-party data it’s in how strategically you build and layer your audiences.
If you look at the interface, you’ll see the core behavioral pillars of Amazon DSP audience creation based on ASINs:
• 𝐏𝐫𝐨𝐝𝐮𝐜𝐭 𝐕𝐢𝐞𝐰𝐬 (Retargeting)
• 𝐏𝐫𝐨𝐝𝐮𝐜𝐭 𝐏𝐮𝐫𝐜𝐡𝐚𝐬𝐞𝐬 (Loyalty/Cross-selling)
• 𝐒𝐢𝐦𝐢𝐥𝐚𝐫 𝐏𝐫𝐨𝐝𝐮𝐜𝐭 𝐕𝐢𝐞𝐰𝐬 (Conquesting/In-Market)
• 𝐏𝐫𝐨𝐝𝐮𝐜𝐭 𝐒𝐞𝐚𝐫𝐜𝐡𝐞𝐬 (High-intent Consideration)
But pulling these levers out of the box isn't enough. Here are my 𝐩𝐞𝐫𝐬𝐨𝐧𝐚𝐥 𝐢𝐧𝐬𝐢𝐠𝐡𝐭𝐬 on how to actually turn these targeting types into a high-ROI strategy:
• The default is often 30 days, but one size 𝑛𝑒𝑣𝑒𝑟 fits all. For low-cost, impulsive buys (like snacks or cosmetics), a 7- to 14-day lookback for Product Views keeps your ad spend tight and highly relevant. For high-ticket items (like electronics or furniture) where the consideration phase is long, stretch that lookback window to 60 or 90 days.
• Want to destroy your ROAS? Keep showing ads for a refrigerator to someone who bought that exact refrigerator yesterday. When setting up a Product Views campaign, always use the Product Purchases behavior as a negative audience exclusion. Protect your budget.
• Targeting "Similar product views" is the ultimate way to capture market share from your competitors. If you know your product has better reviews or a better price point than a top competitor, input their ASINs and target their viewers.
• Users who search for a specific term or product but don't click are incredibly high-intent. Layering Product Searches with tailored creative that highlights a unique selling proposition (USP) or a coupon code is one of the fastest ways to drive incremental conversions.
Amazon DSP gives you the raw data, but your strategy dictates the results. Stop running generic campaigns. Segment by intent, tweak your lookback windows based on buying cycles, and always exclude past buyers unless you're aiming for a subscribe-and-save play.
What’s your go-to audience layering strategy in DSP?
Let’s talk in the comments! 👇