Amazon sells an ad that asks your buyers for a review. My best seller can’t use it, and the one product that can is the one I shouldn’t pay to promote.
Review Requests has one published rule that matters: a product qualifies only under 1,000 ratings and reviews.
I checked my gardening gloves first, because they sell the most. On November 5, 2026, they had 1,819 ratings. Locked out.
The product that qualifies is my espresso cup set, at 351 ratings. I should not pay to promote it. The reason has nothing to do with reviews.
What Amazon has said, and what it hasn’t
Amazon announced Review Requests at unBoxed on September 29, 2026, and opened a US-only beta in late October 2026. It reaches “customers who recently purchased your product” on Amazon’s busiest pages, like the homepage, and asks for a one-tap star rating. The campaign pauses itself at 1,000 ratings. Bids and budgets work like your other campaigns. (Amazon Ads announcement)
Amazon’s one result: in a closed US beta, products got about three times as many ratings and reviews per week. Amazon shared no sample size and no spend.
What Amazon has not said is what you are billed for. Agencies that covered it disagree: per click, per click to submit, per impression. Nobody outside Amazon knows what one new rating costs.
And nothing in it picks happy buyers. A paid prompt asks everyone, so unhappy buyers’ ratings arrive faster too.
What my own store showed
Amazon’s free Request a Review is a button in Seller Central, one per order, and Amazon does not send it for you. Our Feedback Genius sends mine: 11,157 from November 2025 through October 2026. So the paid ad would be a second ask, to buyers who ignored the first.
The cups sell for $25.99. After fees and cost, a sale makes about $2. Over 30 days, from October 5 to November 3, 2026, two buyers returned a set and the product lost $18.57. If a few paid clicks win one rating, that rating can cost more than a sale earns. And ratings would not fix the margin or the returns.
This ad does not make a product better. It makes it louder, and it charges you for the volume.
Five checks before you pay for a review prompt
Five checks, product by product:
- Under 1,000 ratings?
- Is the free ask going out on every order? Run that before paying for a second one.
- Are buyers happy? Read recent reviews and return reasons. Fix the product first.
- Is review count what holds sales back? Paid ratings help a good product, not one that loses money.
- If the ad shows in your ad console, test with a hard cap and judge it by spend per new rating. If it doesn’t show yet, finish checks 1 to 4 so you know which product to test the day it does.
Checks 3 and 4 are the slow ones. With the Seller Labs MCP server connected to Claude, one question pulls profit, returns and recent reviews for every product.
My best seller can’t use this ad. The one that can, I shouldn’t pay for. Not yet.
Paying for reviews only works on a product people already love.
