ReviewProof ⚖

How should you moderate reviews without suppressing them?

By the Ingenious AI team · Published 14 November 2026 · Reviewed 14 November 2026

Run the free review audit See pricing · See a sample provenance report

Short answer

Moderate with three logged states and no delete key: every imported review starts as 'kept (default — no suppression)'; a decision can move it to published or flagged internally, and every decision is written into the provenance chain with its reason. The review itself is never silently deleted. That model keeps your own widget honest, keeps suspected fakes available as evidence, and produces the moderation log — negatives kept included — that an evidence export uses to answer a suppression charge.

What is review moderation?

Review moderation is the set of decisions a business makes about reviews in its own systems: what to publish on its own surfaces, what to flag for investigation, what to dispute through a platform. Suppression-free moderation adds one constraint that changes everything: every state change is recorded, reversible in the record, and the underlying review is retained. The business controls its own publication; it does not control the historical record.

Decide in public-to-yourself: the reason is the moderation

A moderation decision without a logged reason is indistinguishable from tidying. 'Flagged internally — incentive marker; offer traced to a weekend promotion, ended 12 May' is a decision an examiner can follow. The habit to build is that the reason is typed at the moment of decision, in the plainest words that are true, assuming the one reader you must convince is a sceptic with the full file. ReviewProof structures this: moderation events append to the chain of custody, the default is kept, and there is no path in the product that makes a review vanish. Tools shape behaviour; this one is shaped against the quiet fix.

The FTC fake-review rule, 16 CFR Part 465, has been in force since October 2024. It bans fake and AI-generated reviews, buying positive or negative sentiment, insider reviews without disclosure, review suppression, and misrepresenting review sites as independent. ReviewProof was built as the evidence layer for that rule: imported reviews carry a provenance chain — import, transaction link, analysis, moderation decisions — and the Pro evidence export includes the moderation log with negative reviews kept, which is how a business shows suppression did not happen. Proof, in this design, is a record anyone can inspect, not a badge a business awards itself.

Your widget and the platform are different jurisdictions

Confusion here causes most suppression mistakes. On a platform, you do not moderate at all — you reply, and you dispute through the platform's process with evidence. On your own site, you publish: the verified widget shows transaction-linked reviews in kept or published states, which is a provenance rule, not a sentiment rule — a kept one-star verified review displays, and should. Flagged-internal is the honest middle state: excluded from your promotion while investigated, retained in the record, its flag and reasoning visible to anyone auditing you. Three jurisdictions, three behaviours, one log.

After the first audit, the habit that matters is cadence. Re-import on a regular rhythm — monthly suits most locations — so new reviews enter the provenance record while the underlying transactions are still easy to match. Each import logs provenance events, moderation starts from kept (no suppression), and the certificate for any review shows its chain: imported, transaction-linked or not, analysed with its signals, moderated with its decision. A record built gradually like this is worth far more in a dispute than one assembled in a panic the week the dispute arrives.

Mistakes that undo the work

Creating a 'removed' folder outside the system is silent deletion with filing. Moderating by rating (flagging every one-star, keeping every five-star unexamined) reproduces suppression inside a logged tool — the log itself will show the pattern. And bulk decisions ('flag all 40') without per-row or per-cluster reasons write a log that proves process was skipped.

Enforcement and platform action both turn on records, not intentions. A business that can produce the underlying row, its source, its date, its transaction link or the honest absence of one, and every moderation decision taken about it, is in a different position from a business holding only a star average. That is why ReviewProof never silently deletes: imported reviews default to kept, moderation is limited to kept, published or flagged internally, and every decision is written into the chain of custody. The evidence pack (a Pro and Agency feature) assembles that chain so it can be handed over as one document.

Worked example: moderation week at a two-branch retailer

A fictional worked example, written to show the numbers and the decisions in practice.

After its first import of 287 reviews, a retailer's moderation session applied its written rules to 31 flagged rows. Fourteen gating rows: kept, with the source script replaced. Nine incentive rows: flagged internally, promotion ended, reason logged per row. Five suspected bought rows from a burst cluster: flagged internally, platform dispute filed with the evidence pack, rows retained. Three AI-signal rows judged genuine formal writers: kept, with the judgement noted. Total deleted: zero. When the platform later reviewed the dispute, the retailer's moderation log — decisions, reasons, dates — was the document its reviewer quoted in closing the case.

Your five-action checklist

  1. Adopt the three states — kept, published, flagged internally — and no others
  2. Write the reason at the moment of each decision, in plain words
  3. Keep suspected fakes in the record; dispute them through the platform
  4. Apply rules by signal and evidence, never by star rating
  5. Review the moderation log itself each cycle for sentiment patterns

Frequently asked questions

Can a flagged review still be genuine?

Yes — flagging is an internal investigation state, not a finding. That is why the review is retained and the reasoning logged rather than the row removed.

Does flagging a review remove it from our site?

Flagged-internal reviews are held out of your promotional display while investigated; kept or published verified reviews are what the widget shows.

What happens on the platform itself?

You cannot moderate there. You reply publicly and, for fabricated or policy-breaking reviews, dispute through the platform's process — the evidence pack is your supporting document.

Why is there no delete in ReviewProof?

Because a silent-delete path would break the chain of custody and recreate suppression inside the tool. Retention with logged states is the design's core commitment.

Keep reading

Related guides from the ReviewProof library:

Start with the tool itself: open the free audit desk, see how ReviewProof works, compare Free, Pro and Agency plans, inspect the sample provenance report, or browse all ReviewProof guides.