What commerce reviews operations means for retail feedback teams

What to take away

Why have review operations changed?

The direction of travel is clear. The Digital Markets, Competition and Consumers Act 2024 gives ministers powers to set rules on fake and incentivised reviews. The government signalled in 2024 that it intends to use them. The Competition and Markets Authority (CMA) is the body that will enforce the new regime, and it already enforces the current rules. Enforcement dates have moved, so treat any timetable as provisional and build the controls now.

The older rules still apply. Under Regulation 5 of the Consumer Protection from Unfair Trading Regulations 2008, a commercial practice that misleads the average consumer about reviews is a misleading action. Those Regulations are enforced by the Competition and Markets Authority and by local authority Trading Standards. Publishing only five-star feedback is not automatically unlawful, but suppressing negative reviews can be.

What this really changes is ownership. A review function without a named accountable person cannot answer the CMA, a platform or a journalist. It also cannot prove that its moderation decisions were consistent.

Who owns commerce reviews operations?

Ownership should sit with whoever already controls the customer record, usually the ecommerce or customer operations lead, with a second sign-off from legal or compliance. Marketing can run the request programme, but it should not control what gets published.

A simple split works. One person owns the collection flow and the request schedule. A second owns moderation decisions and the escalation log. A third owns reporting, including response rates and dispute volumes. Where teams are small, two people can cover three roles, but the same person should not both solicit reviews and decide which ones appear.

That separation matters because the most common failure is not fraud. It is a stretched team approving everything in a queue at 6pm on a Friday. Every escalation should also have a named route, covering who is told, within what time, and what they may decide without a meeting.

Mapping responsibilities properly is the subject of commerce reviews team roles: data and sources, which is worth reading before you write a job description.

How should you collect reviews without breaching marketing rules?

Collecting a review is a marketing contact. Direct marketing by email and text is governed by the Privacy and Electronic Communications Regulations 2003 (PECR), which the ICO enforces. The ICO direct marketing checklist sets out the consent, soft opt-in and opt-out tests you need to satisfy. It applies to a post-purchase email just as much as to a newsletter.

Three habits keep you clean. Ask once, with a clear opt-out in the same message. Keep the request tied to a transaction so the soft opt-in argument holds. Never make a review a condition of a refund, a warranty or a discount on a future order.

Suppression is the harder question. If you filter requests to avoid unhappy customers, you are shaping the published set. A defensible approach is to invite everyone and then moderate the content, not the audience. For worked examples of what that looks like in a queue, see commerce reviews quality checklist in England.

What does moderation actually involve?

Moderation has four jobs: remove content that is unlawful or off-topic, verify that the reviewer bought something, flag incentives, and log the reason for every decision. Anything beyond that is editorial judgement and should be documented as such.

Write your rules down before you need them. A short moderation policy covering profanity, personal data, competitor mentions and medical claims will settle most arguments. The test is whether two reviewers reading the same policy would reach the same decision on the same review.

A review record is personal data. Holding it, and publishing a reviewer's name or profile, falls under the UK GDPR and the Data Protection Act 2018, which the Information Commissioner's Office regulates. Most retailers rely on legitimate interests as the lawful basis for publishing, and set a retention period for rejected and removed content in the same policy.

Give customers a route to challenge a decision. An appeal address that reaches a person, not a form, reduces the chance that a disappointed reviewer takes the complaint to the platform's own dispute process instead. Log a reference number for every escalation so that a run of appeals about the same product or the same moderator becomes visible.

Speed matters commercially. A review queue that runs a week behind will be publishing feedback about a product version you no longer sell. Set a target, publish it internally, and measure against it. For a fuller operating sequence, commerce reviews operating workflow in England sets out the steps in order.

Decision table: matching the approach to the situation

Whichever row of that table fits, the platform you pick matters more than the volume it promises. The UK market includes Trustpilot, Feefo and Reviews.io for hosted collection, and enterprise options such as Bazaarvoice and Yotpo that plug into a retailer's own stack. Ask each shortlisted platform where review text is rendered, whether the full moderation log can be exported, and who owns the review data if you leave.

How do you keep review pages visible to search?

Review content is often rendered by JavaScript, which means it may never reach the index. Google's overview of its crawlers and fetchers explains how Googlebot discovers and requests pages. It is the reference point for deciding whether your review widget is crawlable.

The operational fix is mundane. Confirm that review text appears in the initial HTML response, or that your platform serves a prerendered version. Then check that paginated review pages are not blocked by robots rules left over from a staging environment.

Structure also matters. Use one canonical page per product for reviews rather than fragmenting feedback across campaign landing pages. If your reviews live on a subdomain owned by a third party, you are relying on their indexing decisions, not yours.

Markup is a separate decision from crawlability. Structured data should describe reviews a visitor can actually see on the page. Adding markup for hidden or aggregated feedback creates a search visibility risk that sits alongside the regulatory one.

How do you advertise with review content?

Paid distribution of UGC needs its own permissions. A customer who agreed to a review has not agreed to appear in an advert, and the two consents are not interchangeable. Keep a separate rights record with the wording the customer actually saw.

Platform setup is the easy part. LinkedIn's getting started guide for advertisers covers campaign structure and targeting. The same discipline applies on any channel where you plan to put customer quotes in front of a new audience.

The harder part is substantiation. In the UK, the Advertising Standards Authority applies the CAP Code, which requires advertisers to hold evidence for the claims they make. If an advert claims a rating drawn from a set number of customers, you need the source, the date range and the population. For example, a team running a paid campaign on a 4.8 average from 2,000 reviews collected over eighteen months should be ready to explain both numbers if asked.

What service standards should you set?

Standards turn good intentions into something you can audit. Four are worth publishing internally: time to first moderation decision, share of reviews with a verified purchase flag, response rate to reviews raising a service issue, and the lag between publication and reporting.

A workable baseline for a mid-sized UK retailer is a first decision within two working days. A verified purchase flag should sit on at least 90 per cent of published reviews, and any review that raises a complaint should get a reply within three working days. Those are illustrative targets rather than benchmarks, so set yours from your own history.

Review the standards quarterly. If your queue is consistently clear within a day, your target is too loose. If you breach it every December, the rota is wrong, not the standard. Commerce reviews service standards in England covers how to set and evidence them.

What does 2027 look like for review teams?

Expect three pressures. Regulators will want evidence of process rather than assurances of good faith. Platforms will keep changing what review data they expose and how. Customers will keep testing the boundary between feedback and complaint, especially where a review is the fastest route to a human.

Budget follows risk. Teams that can show a documented policy, a decision log and a named owner will find it easier to justify software spend than teams that cannot. That is a change from the last decade, when review tooling was sold mainly on volume and sentiment dashboards.

The skills mix shifts too. The CIM customer experience training route is one way to build the feedback-handling side of the role. It helps particularly where review content feeds into wider service improvement rather than sitting in a marketing silo.

If you plan headcount or tooling for the next financial year, commerce reviews: trends and outlook for England in 2027 sets out the changes worth building into the plan.

Where does reporting fit?

Reporting exists to answer three questions: are we collecting enough, are we moderating consistently, and are we responding to problems. Anything that does not serve one of those is decoration.

Keep the report short. Volume collected, volume published, volume rejected with reasons, median time to decision, and the count of reviews mentioning a service failure. Add a comparison against the previous period and stop there.

One caution. Aggregate star ratings hide distribution. A 4.5 average built from mostly fives and a cluster of ones tells you something a single number cannot. It is usually the cluster that predicts returns and contact volumes.

How do you handle a fake review allegation?

Treat it as an incident with a clock on it. Preserve the review, the account history and the moderation log before anything is removed. Then decide whether the content is fake, incentivised or simply unwelcome, because the three require different responses.

If the review is fake, remove it and record why. If it is genuine but damaging, publish it and respond. If it is incentivised, check whether the incentive was disclosed. An undisclosed incentive is the more serious problem under the misleading actions test.

Report coordinated fake review activity to the platform and keep the reference. Complaints about misleading reviews can also go to the Competition and Markets Authority, or to Trading Standards through the Citizens Advice consumer service. If the same pattern repeats, it belongs in your risk register, not your inbox.

Common questions

These are the questions that come up most often when a retail team takes on review operations.

Do I need consent to email a customer asking for a review?

Usually the soft opt-in covers it. The test itself sits in the Privacy and Electronic Communications Regulations 2003 (PECR), and the ICO's direct marketing checklist sets out how to apply it: you collected the address during a sale, the review request relates to similar goods or services, and you offered an opt-out at collection and in every message.

Can I delete a negative review?

Only if it breaches your published moderation policy, for example because it contains personal data, abuse or content unrelated to the purchase. Removing a genuine negative review to protect an average rating risks being a misleading action under Regulation 5.

How quickly should a review go live?

There is no legal deadline, but a first moderation decision within two working days is a sensible internal target for a UK retailer. Slower queues mean you publish feedback about products and processes that have already changed.

Who should sign off the moderation policy?

Someone with authority over both the customer record and legal risk, typically the ecommerce or customer operations lead, countersigned by legal or compliance. Marketing can draft it but should not be the sole approver of what gets published.