Stop relying on ROAS as your north star for advertising on Amazon when you can u
By Jasonlandro · October 7, 2026 · Curated by George's Blog
Stop relying on ROAS as your north star for advertising on Amazon when you can use customer lifetime value
ROAS only gives you the return you get on a search term, which could be branded (or not) and isn’t necessarily incremental (you might have gotten the sale anyway)
It also doesn’t account for your customer acquisition cost, customer lifetime value, brand lift, organic rank on Amazon, and more
Instead, you should anchor your ads strategy to the return specific customer segments have generated across products in your catalog over months or years
If your measurement systems weren't built to look beyond a single conversion window or the standard attribution period for an advertising conversion, that can be an impossible question to answer
The challenge is particularly acute for brands selling durable goods
If your product has a 2-year repurchase cycle, short-term attribution windows don’t just undercount your ads but structurally undervalue the campaigns that find you the most relevant customers
What we've seen shift with clients using Amazon Marketing Cloud is that brands can measure beyond the default window and beyond Amazon’s owned channels
You can analyze new-to-brand acquisition over a time horizon that actually reflects your categorys purchase cycle, identify which audience segments generate the most long-term value across ASINs and categories, and connect off-Amazon conversions through Amazon Ad Tag and Conversion API
Specifically, you can upload your own offline data from other retailers or your D2C site to understand how your advertising is driving conversions across all channels, not just one
Thats a fundamentally different picture than "who clicked and converted this week?"
The brands I see winning here arent just measuring more
They're measuring longer and more accurately
As a result, the budget decisions that follow often look very different because these brands are looking at the full picture, not part of it