Strategy
Part of A reputation management strategy for disputed reviews
Reputation management strategy mistakes to avoid
Reputation management strategy mistakes that break review integrity, from removal targets to unsupported impact claims — and the fix for each.
What to take away
Seven reputation management strategy mistakes recur in review handling, in testimonial use and in impact claims. Each one has a procedural fix, and each fix should leave an audit trail.
- A plan that rewards fewer negative reviews pushes a team toward selective suppression. Replace that target with a traceable investigation and response process.
- Label every item as opinion, disputed fact, verified record or unknown, so a later correction stays possible.
- Service messages and promotional messages need separate classifications, each with its own evidence and suppression check.
- Urgency is not a reason to share credentials. Record who can draft, approve, publish, revoke and recover each account.
- A movement in rating or ranking does not show your work caused it. A credible counterfactual does.
Why removal targets fail review integrity checks
The first mistake is treating removal as the goal. A team measured on fewer negative reviews will chase deletions, dispute honest complaints and bury awkward feedback. The CMA's guidance for online review sites says genuine negative reviews should not be withheld merely because they are unfavourable.
That guidance sits inside a wider regime. The Digital Markets, Competition and Consumers Act 2024 gives the CMA direct powers over fake and misleading reviews. Suppression belongs on the same risk register as fabrication.
The correction is procedural. Replace fewer negatives with a traceable investigation and response process: log the item, test it against your own records, answer it in public, record the outcome. One England operating log can carry that for a whole group, since the rules cited here apply UK-wide.
Set the log where the reviews arrive. Google Business Profile, Trustpilot, Feefo and Reviews.io each keep their own thread and their own timestamp, so the same four fields — source, claim, status label and outcome — need to be mirrored in whichever reputation-management dashboard the team works from. A tool that cannot show the status label beside the original review will not support a correction later.
Until that process exists, the release gate stays closed: an empty reputation dossier — no logged items, no status labels and no named approver — is not signed off, because nothing in it can evidence a decision.
Example: one row from a review-response log
A worked row shows what a traceable record holds when a complaint is disputed.
Review-response log row
| Field | Example entry |
|---|---|
| Source and date | Google Business Profile, 14 May |
| Claim | Fitter left a leaking radiator and will not return |
| Status | Disputed fact, customer account not yet reconciled with the job sheet |
| Evidence ID | JOB-4471, photographs and signed handover sheet |
| Owner and action | Service lead, call booked within two working days |
| Public reply | Acknowledges the complaint, offers a named contact, admits nothing |
| Recheck | 28 May, outcome recorded before the item closes |
The row is the unit of account. A removal target cannot be reconciled against it, and neither can a meeting note.
Sorting disputed claims from service messages and promotions
The second mistake is the unlabelled claim. Copying an allegation into an internal profile without a status label makes later correction hard. The ICO's accuracy guidance in its Guide to the UK GDPR asks for clarity about the source and status of personal data. Record opinion, disputed fact, verified record and unknown as four separate labels.
The third mistake is mixing service and promotional messages. A reply about a complaint turns into marketing when it carries promotional content. The ICO's direct marketing guidance applies a purpose, content and context test, so classify service and promotional messages separately. Each needs its own evidence and suppression checks.
Review-led adverts carry a further layer of rules. A rating or testimonial in paid or owned media is itself a marketing communication, and the advertiser carries the burden of holding evidence for it.
Testimonials, access and account security before release
The fourth mistake is treating authenticity as substantiation. CAP's current testimonial advice, in Section 3 of the CAP Code on misleading advertising, requires documentary evidence and warns that a testimonial claim must not mislead. Check the exact quotation, the permission, the relationship, the product context and the final placement.
The fifth mistake is unprotected shared access. Shared credentials make a correction harder to contain. The NCSC's guidance on social media, published as 'Social media: how to use it safely', recommends measures including two-step verification and controlled shared access. Record who can draft, approve, publish, revoke and recover each account.
The sixth mistake is leaving accessibility testing until the end. The GOV.UK Service Manual's 'Testing for accessibility' guidance explains why automated checks need manual and assistive-technology testing. Set the private contact route, the public reply, the challenge journey and the appeal path as test tasks before approval.
If a review breaches a platform's rules, use the platform's own route first. Flag it in Trustpilot, Feefo or Reviews.io, or report it in Google Business Profile, then appeal through the same case if the decision goes against you, and keep the case reference. Where a business appears to suppress genuine negative reviews or publish fake ones, report it to the CMA through its contact route on GOV.UK; individual disputes with a trader go to the Citizens Advice consumer service, which passes them to Trading Standards.
Proving the work caused the change
The seventh mistake is claiming causation from a movement in rating or ranking. A later rating or search-position movement can have many causes. HM Treasury's quality guidance for impact evaluation, the Magenta Book, explains that monitoring change does not establish that an intervention caused it, and it asks for a credible counterfactual. The causal distinction holds for commercial reputation claims as well.
Pick the outcome unit before you start — a resolved case, a corrected listing or a closed complaint. Our comparison of attribution methods holds several approaches against the same unit, which is the part most plans skip.
Correcting the planning record
- Replace removal targets with a traceable investigation and response process.
- Label every item as opinion, disputed fact, verified record or unknown.
- Classify service and promotional messages separately, each with its own evidence.
- Check each testimonial quotation, permission, relationship and placement.
- Test the private contact route, public reply and challenge journey before approval.
- Record who can draft, approve, publish, revoke and recover each account.
Retest the corrected path with invented material before it handles a live case. Preserve the failed and revised versions, who approved the change and what would reverse it. If the team cannot reproduce the fix, the item stays open.
Common questions
Why does the removal target survive so long?
It is easy to count and it looks like a clean result. An investigation process is harder to measure, so it loses the argument. Measure resolution instead, meaning how many items were investigated, answered and closed with evidence attached.
What if a review states a fact we can disprove?
Put the evidence beside the claim rather than deleting the claim. A dated job sheet, photograph or signature can support a short factual reply. Do not publish the reviewer's personal data to make the point.
What must survive after a correction goes out?
Keep the failed version, the revised version, the approver's name, the evidence ID and the trigger that would reverse the change. Under ICO accountability guidance you should be able to explain how reviewer personal data was handled at each step.
