G2, Capterra and Trustpilot Programmes That Pass Audit

Grid inclusion in a G2 category needs ten approved reviews from the previous twelve months, and that is the number most programmes are built around. It is the wrong number. Trustpilot's TrustScore is a time-weighted Bayesian average, so forty reviews collected in one campaign week are worth progressively less every month afterwards while a competitor collecting four a week passes you inside a year. We design for the weekly rate. The campaign burst is the thing we are usually hired to undo. The collection trigger is the compliance surface: Since the United States rule on fake and manipulated reviews took effect in October 2024, and under the EU Omnibus Directive before it, the prohibited parts are specific: incentives conditional on what the review says, suppressing negative submissions, screening customers through a satisfaction question before deciding who gets the invitation, and publishing without disclosing how reviews were collected. Nearly every programme we inherit is gated, and almost always by accident, because a marketing automation flow was set to send the request only to promoters. We rebuild the trigger on a product event instead: a renewal, a completed onboarding milestone, a support ticket closed at any outcome. The sample stops being self-selecting, the average drops by two or three tenths of a point in the first quarter, and the profile starts reading as real to buyers who have learned to distrust a wall of fives. Invitations are written per platform, never once: G2 permits incentives inside its own gift card programme, disclosed, offered to every invitee regardless of what they write. Capterra's rules differ. Trustpilot's automatic invitation model is built precisely to stop cherry-picking and behaves badly when handled as a list upload. A single invitation template violates at least one of the three, so we write three, and we write the disclosure line that has to sit inside them. Syndication means one complaint lands three times: Capterra, GetApp and Software Advice share the Gartner Digital Markets review pool. An unresolved complaint appears across all three with three separate search footprints, and the response you write is the one buyers read in each. We treat the pool as a single surface and prioritise responses accordingly. The response line that gets clinics in trouble: In health and wellness, a public reply confirming that a reviewer was a patient discloses special category data under Article 9 of the GDPR, and the fact that the person disclosed it first does not transfer the lawful basis to you. Replies we write acknowledge the review, offer a named offline route, and assert nothing about whether the person was treated, when, or by whom. We have read published enforcement decisions against practices that got this wrong on a public review page, and it is not a theoretical risk. What we refuse: Writing reviews. Buying them. Gating by sentiment in any form, including the polite version where a satisfaction score decides who receives the link. Incentives that vary by rating. Mass-disputing accurate negative reviews to see what the moderators remove. If your current agency does any of these, say so at the start, because the remediation is a different piece of work from the build. Not included: Platform licence fees and paid profile placement, which you buy directly and which we have no financial interest in. Case study production. Analyst relations, which is a separate discipline and a separate budget. What gets measured: Weekly collection rate against target, the twelve-month rolling count per category, response latency on reviews under three stars, and the trailing average with and without the incentive cohort separated out, because you should know which of your two numbers is the real one.