Stars & review count — the trust row on the SERP tile.

Directly below the price row sits the rating row: five stars (some hollow, some filled, some half-filled) and a number in parentheses. Small, easy to overlook when you're building a listing. For shoppers, it is often the deciding signal between two otherwise comparable tiles — and the threshold that determines whether your listing even appears in the result set when filters are applied.

How the star rating is actually computed

The average shown next to your title is not a plain arithmetic mean of every star ever given. Amazon's algorithm weights ratings on at least three dimensions:

• Recency. A 5-star rating from last month counts for more than a 5-star rating from three years ago. This is why a listing that briefly shipped a defect can recover its visible average within months of fixing the issue.
• Verified purchase. Ratings tied to a confirmed Amazon purchase outweigh unverified ones, sometimes substantially. Unverified reviews still display but contribute less to the headline number.
• Helpfulness signals. Reviews with photos, longer text, and "helpful" votes from other shoppers carry more weight than a one-tap star with no comment.

The result is a number rounded to one decimal (e.g. 4.3) that Amazon then renders as a half-star icon on the tile. The same underlying number — not the rounded display — drives the "4 stars & up" filter in the left rail, which is why slipping from 4.0 to 3.9 has an outsized impact on qualified traffic. The visible change is one pixel of star fill; the invisible change is being filtered out of every search where a shopper ticks the 4-star checkbox.

What the count actually counts

The number next to the stars used to be the review count: how many people had written a written review. That changed in 2019 when Amazon rolled out one-tap star ratings. Today, in most marketplaces, the tile shows ratings, which includes:

• Written reviews (with or without photos).
• One-tap star ratings submitted from order history or post-purchase emails, with no written content attached.

The ratings count is usually 5× to 20× larger than the strict review count, depending on category. For the shopper, "4.4 stars · 2,847 ratings" reads as trustworthy; "4.7 stars · 12 ratings" reads as promising but unproven. The transition from "new" to "established" in most categories sits somewhere around 50–100 ratings — below that, even a high average struggles to convert against a 4.2-star competitor with 3,000 ratings.

The two cliffs on the filter rail

Two thresholds in the left-rail filter change visibility itself, not just appearance:

• 4 stars & up. The most-ticked review filter on the SERP. Slipping from 4.0 to 3.9 removes you entirely from every search where a shopper applies it. Internal Amazon data and third-party studies routinely put this at 30–50% of qualified searches in high-consideration categories like beauty, supplements, and electronics.
• 3 stars & up. A smaller second cliff. Below 3.0 you disappear from almost every filtered search and are typically priced into clearance territory regardless.

Defending the 4-star line is the single most important review-side job for an established listing. A repeated 1-star pattern (a defect, a sizing issue, a packaging problem) is not just a customer-service issue — it is a SERP-visibility issue.

Stages of review growth — what to do at each

Different stages of the rating curve call for very different tactics:

1. 0–10 ratings. Survival mode. Enrol in Amazon Vine (Brand Registry required) to get up to 30 honest, professional reviewers; budget around €170–€200 per Vine product. Send every eligible buyer a Request a Review tap. Never solicit positive reviews specifically — that's a policy violation.
2. 11–50 ratings. Defend the average aggressively. A single 1-star at 4.0 average with 12 ratings can shift the visible number from 4.0 to 3.8 overnight. Read every negative review for product issues and fix them in the next production batch.
3. 51–500 ratings. Scaling phase. Use the Manage Your Customer Engagement tool, the Request a Review button in Seller Central, and post-purchase emails sent via Amazon's permitted channels. The average is more resilient now — a single 1-star moves the needle far less.
4. 500+ ratings. Optimisation phase. Moving the average from 4.2 to 4.3 takes months. Focus on review content: surfacing photo reviews, getting "top reviewer" coverage, and on returns / quality issues that drive the new 1- and 2-star negatives that drag the recency-weighted average down.
5. 5,000+ ratings. Defence phase. The count itself is a moat — competitors can't catch up in any reasonable time. Protect it with consistent product quality, proactive engagement with negative reviewers, and listing-side defence against ASIN-hijack attempts that would split your reviews across a new ASIN.

What you cannot do — and what triggers suspension

The following are explicit Amazon policy violations and a leading cause of seller-account suspensions:

• Incentivised reviews (offering a discount, refund, or free product in exchange for a review).
• Reviews from friends, family, or anyone with a personal relationship to the seller.
• Review-trading schemes, Facebook groups, and any third-party "review service" that promises reviews for payment.
• Asking specifically for a positive review or 5-star rating, in any channel — only neutral requests are allowed.
• Inserting product cards into the package asking for a review in exchange for warranty or support access.

The only safe sources of new reviews are: Amazon Vine, the Request a Review button, the Manage Your Customer Engagement tool, and emails sent through Amazon's Buyer-Seller Messaging service using neutral, non-incentivised language.

Seller vs Vendor — who owns the response

Reviews live on the ASIN, not on the seller account, so both Sellers and Vendors see the same review wall on the detail page. The operational difference is who can respond. Brand-registered Sellers can respond publicly to negative reviews via the Brand-Registered Comment feature. Vendors typically route review responses through their Vendor Manager and have slower turnaround. In both cases, response language has to follow Amazon's community guidelines — no discount offers, no off-Amazon contact, no disputing the reviewer's experience publicly.

What to take into the next episode

The rating row finishes the tile. We've now walked every zone of the SERP — anatomy, search bar, filters, sponsored placements, organic positions, badges, image, title, price, reviews. The last episode of Module 2 turns that walk into something actionable: a one-page listing brief that captures every SERP-side decision before any creative work begins.