Why reviews convert
A product page makes claims. A review tests them. Shoppers know the store wrote the description and chose the photographs, so the description can only take them so far; a stranger reporting that the jacket runs small, that the blender is loud, that the colour is closer to navy than the photo suggests, gives them the information they need to decide. The effect on conversion rate (CVRShare of visits or clicks that convert.) is strongest where uncertainty is highest: unfamiliar brands, products that are hard to assess from images, categories with a history of disappointment, and higher-priced items where a mistake costs more.
Reviews also do a second job that is easy to miss. They carry language. The words customers use to describe a product — what they bought it for, what surprised them, what they compared it with — are the words other shoppers search for and the words a product page should use. A store that reads its reviews learns how its products are understood, which feeds product copy, advertising and the questions the page should answer up front. The trust and proof guide covers the full set of proof elements a page can carry — guarantees, press, certifications, counts — and how to arrange them; this page concentrates on the one proof element that a store has to earn one customer at a time.
Collecting reviews
Most stores have far fewer reviews than their order volume could support, because asking is left to chance. A programme replaces chance with a consistent request after every eligible order, and the table sets out the ways reviews arrive and what each needs.
| Source | How it works | What it needs |
|---|---|---|
| Post-purchase email or SMS request | Automated message after delivery asking for a rating and comment | A trigger tied to delivery, not dispatch; a one-click rating that leads to the form |
| On-site prompt | Request shown to signed-in customers who have received an order | Account login or a verified-purchase link |
| Review platform syndication | Third-party review services that collect and distribute reviews | A contract; the platform's rules on display and moderation |
| Marketplace reviews | Reviews left on marketplaces where the product is also sold | Cannot usually be imported; worth monitoring for product insight |
| Photo and video reviews | Customers upload images alongside text | A clear ask, easy upload, and a visible place where photos appear |
| Customer questions and answers | Shoppers ask, previous buyers or staff answer | Routing to people who will answer quickly |
| Imported reviews from a previous platform | Migration when changing review tools | Verified-purchase status preserved; dates kept honest |
Two principles run through the table. Every review should be traceable to a purchase, because verified-purchase status is what makes the collection defensible and what shoppers increasingly look for. And the request should go to every customer, not to the ones the store expects to be pleased, because selective asking is a form of manipulation that regulators have started to name explicitly.
Timing and the request
The request lands best after the customer has used the product, which is later than most stores send it. A message sent on dispatch asks for an opinion about something the customer has not yet seen. A message sent on delivery is better. A message sent after enough time for the product to be used — a few days for clothing, longer for anything that needs installing or a period of use — produces reviews with substance, because the customer has something to say.
The request itself should be short, personal to the order, and specific. Name the product, show its image, ask one question that invites detail rather than a bare star rating, and make the first click do the rating so the form appears already half-complete. One reminder a week or so later, and then stop; repeated requests irritate and look like pressure. Where the store sends by SMS, the consent and frequency rules for messaging apply, and the e-commerce email and SMS flows guide places the review requestThe message or prompt that asks a customer to leave a review, sent once the product has been used and offered to every customer on the same terms. in the post-purchase flow where it belongs alongside delivery updates and care instructions.
Avoid asking for a review in the same message as a discount or a loyalty prompt. The customer should feel asked for an opinion, not negotiated with, and mixing the two confuses the record if the store later has to show that reviews were not conditional on anything. Before a request starts or changes, the review request launch checklist checks who is asked, the timing, any incentive and the path to the form in one pass.
Displaying ratings, photos and questions
Collecting reviews is wasted if they sit on a tab nobody opens. The display decisions are where reviews earn their conversion effect.
The rating near the price. A star summary and review count beside the product title, linked to the reviews section, is the convention shoppers expect. It answers whether people like the product before they read anything else. The product page optimisation guide treats this as part of the page's trust cluster near the buying action.
The reviews themselves, sorted usefully. Most recent first by default, with options to sort by rating and to filter by attribute (size bought, skin type, use case). A summary of recurring themes — runs small, quieter than expected — helps shoppers who will not read twenty reviews.
Photos and video. Customer images are the most persuasive content a store can show because they are unposed. Give them a gallery of their own near the product images, not only inside individual reviews, and let shoppers open the review behind each image.
Customer questions. A questions-and-answers block catches the objections the description missed. Answer quickly, from staff where previous buyers have not, and feed the recurring questions back into the product copy so that the question stops being asked.
Proof beyond the product page. Category pages, search results, the basket and even advertising can carry ratings, and the user-generated content guide covers turning customer photos and posts into social and ad creative with the right permissions.
A footwear store rebuilds its review display
A footwear store has 2,300 reviews across its catalogue but shows them only under a tab labelled "Reviews" at the bottom of each product page, which analytics show 9% of visitors open. The team moves a star summary with the review count next to the price, adds a fit-feedback bar (runs small, true to size, runs large) built from a question in the review form, and creates a customer-photo strip under the product images. They also switch the request trigger from dispatch to seven days after delivery and add a photo upload step. Over three months, the review submission rate rises from 2.1% of orders to 5.4%, roughly a third of new reviews include a photo, and product-page conversion rises by 11% with returns for sizing reasons down by a fifth. The fit bar, not the star rating, is what visitors interact with most.
Negative reviews and responses
Negative reviews are not a problem to be managed away; they are part of what makes the positive ones believable. A page with only top-rated reviews reads as filtered, and shoppers discount it. A page with a few critical reviews, each answered calmly, reads as honest. The critical reviews also tell shoppers what the product's limits are, which reduces returns from people for whom it would not have worked.
The response is where the store shows its character. Reply to critical reviews publicly, quickly and without defensiveness: acknowledge the problem, say what the store did or will do, and take anything personal to a private channel. Never argue, and never ask the customer to change the rating. Where a review describes a genuine product fault, say so and fix the product; where it describes a misunderstanding, correct the product page so that the next customer is not misled.
Read together, praise and criticism are also raw material for advertising: the guide to finding creative angles in customer evidence codes reviews by situation, problem, outcome, objection and proof to find angles for Meta ads.
Moderation should be for abuse, spam and personal data only, with the policy published. Removing reviews because they are negative, delaying their publication while positive ones go live, or displaying an average that excludes them is the kind of practice regulators describe as deceptive, and it is also visible to shoppers who notice when a popular product has no critical reviews at all.
Incentives, fake reviews and the rules
The rules on reviews have tightened in most markets, and the direction of travel is consistent. This section describes the general principles as the FTC's business guidance on advertising and endorsements sets them out; it is not legal advice, and a store trading in several countries should check the position in each.
The principles, stated generally:
- Fake reviewsA review not based on a genuine experience, such as one written by the business, paid for without the experience or bought from a service. are prohibited. Writing reviews for your own products, paying for reviews that were not genuinely experienced, and buying reviews from services are out. The same applies to fake negative reviews of competitors.
- Incentives cannot be conditional on sentiment. A store may offer something for leaving a review, but not for leaving a positive one, and the incentive has to be offered regardless of what the customer writes. Reviews that were incentivised need to be clearly identified as such wherever they appear.
- Suppression is treated like fabrication. Choosing which reviews to publish by rating, or asking only likely-positive customers, misrepresents the overall picture in the same way as inventing reviews would.
- Endorsements from staff, family and affiliates need disclosure. A review from someone with a material connection to the store must say so.
- Company-controlled review sites are a problem. Running a site that presents itself as independent while reviewing your own products is covered by the same guidance.
Beyond the rules, the practical argument is the same: a programme built on genuine, unfiltered, properly disclosed reviews converts because it is believable, and a programme that bends the rules is one investigation or one journalist away from losing the proof it was built on. Regulators in other jurisdictions take a broadly similar line, and consumer-protection enforcement on reviews has become more active, so the guidance linked in the sources is a starting point rather than the whole picture. The same rules apply when AI tools are involved: a generated review is a fake review, and the guide to AI claims and consumer protection covers where AI-assisted marketing most often runs into them. Countdowns, low-stock messagesA persuasion device that tells buyers stock, places or time are limited, prompting a decision; it must be true to be fair to customers., pre-ticked boxes and late fees raise the same question of truth, and the guide to using urgency, scarcity and defaults honestly applies it to each, with the guardrails to test them against.
For sellers whose reviews sit on a marketplace listing rather than their own store, the marketplace channel page for online sales covers earning ratings through the platform's permitted request tools alongside listing and bidding tactics.
Review markup and rich results
Review and rating data can be expressed as structured data (Schema MarkupStructured data vocabulary.) on product pages so that search engines understand the rating, the count and the individual reviews. Where Google chooses to show it, the rating can appear in the search result as a rich result, which affects the click before the page is seen.
The conditions are documented and worth reading in full rather than assuming. Google's structured data documentation for review snippets requires that the rating refers to the product on the page, that the reviews in markup are the reviews visible to users, and that the data is sourced from the site rather than aggregated from elsewhere; self-serving ratings for the business itself are not eligible for the same treatment. Rich results are never guaranteed, and markup that misrepresents what is on the page risks a manual action rather than a rich result. The structured data guide covers implementation, testing and the product markup that review data sits inside.
Reviews also feed the broader quality signals that Google's rater guidelines describe through the idea of experience, expertise, authority and trust (E-E-A-TExperience, Expertise, Authoritativeness, Trust.). First-hand accounts from customers are a form of experience the page could not otherwise show, and a product page with substantive, visible, honest reviews supports the trust part of that picture better than any amount of description. The e-commerce conversion hub sets reviews alongside the other conversion levers a store controls.
Whether reviews earn more beside the add-to-basket button than lower on the page can be tested with the social proof placement experiment.
When orders keep arriving but reviews do not, the troubleshooting page for customers not leaving product reviews checks timing, effort, the flow, the insert card, the rules and the display in turn. When the store is one of several places customers read about the business, the customer reviews and reputation audit scorecard rates asking, display, integrity and replies across every platform together and ranks what to fix first.
Common mistakes
- Asking on dispatch. The customer has not used the product; the review will be thin or absent. Trigger on delivery plus a sensible period of use.
- Asking only happy customers. Selective requests misrepresent the picture and are treated by regulators as a form of manipulation.
- Hiding reviews under a tab. The rating near the price and a photo strip under the images are where reviews change decisions.
- Treating negative reviews as the enemy. They make the positive reviews credible and tell shoppers what the product is not for; answer them well.
- Incentives tied to sentiment. Offering a reward for a good review, or disclosing nothing about incentives, breaks the rules and the trust at once.
- Markup that does not match the page. Review structured data must describe what the visitor can see; anything else risks losing rich results altogether.
- Never reading the reviews. The language, recurring questions and complaints are product and copy research the store has already paid for.
