What a filtered funnel actually is
Most reputation platforms sell the same mechanic under a friendly name. Smart routing. Experience filter. Review shield. The customer is asked how their visit went before they are asked for a review. If they answer well, they go to Google. If they answer badly, they get a private feedback form instead, and Google never hears from them.
It is easy to see why it sells. It feels like protection. And it is extremely common — one Australian platform currently advertises it as a headline benefit on its homepage: “Happy customers go directly to Google Reviews. Negative feedback is routed privately.”
The problem is that it does not work, and that it is not legal in the way people assume it is.
It does not work
There is exactly one published dataset comparing gated and ungated funnels at scale, and it is not flattering. When GatherUp removed gating from its own platform, it measured roughly 10,000 locations across the year before and the year after.
| Google reviews | Filtered | Ask everyone |
|---|---|---|
| Total collected | 32,689 | 53,790 |
| Average rating | 4.66 | 4.59 |
Seven hundredths of a star, in exchange for 21,101 reviews. And the trade is worse than it looks, because 4.59 is the better number. Northwestern’s research across 40+ categories found purchase likelihood peaks between 4.2 and 4.7 stars and then declines as you approach 5.0 — a spotless rating reads as fake. 82% of shoppers go looking for the negative reviews on purpose.
So filtering costs you the review volume that drives your ranking, in order to move your rating in the direction that reduces conversion. It is a bad trade twice over.
And it is not as safe as people think
Google’s policy names it. In plain language, under Rating Manipulation, merchants may not “discourage or prohibit negative reviews, or selectively solicit positive reviews from customers.” There is no ambiguity to argue about.
We should be honest about enforcement: we found no documented case of Google suspending a business specifically for gating. Their detection is aimed at fake and incentivised reviews, and a gated funnel produces genuine reviews from real customers. So the practical Google risk today is low — it is simply entirely at Google’s discretion, and requires no rule change to start.
Australia is where it actually bites, and the precedent is close to exact.
Meriton did not delete a single review. It altered or withheld guest email addresses so that guests it expected to be unhappy never received the invitation. The Federal Court held that misleading, and the penalty was $3 million. Withholding the invitation from predicted-unhappy customers is precisely what a sentiment gate does.
There is no Australian judgment yet applying Meriton to a commercial gating product, and we will not pretend otherwise — that is an argument, not a citation. But the ACCC has already examined these platforms, and its 2023 sweep put the software vendors themselves on notice for aiding, abetting or being knowingly concerned in their clients’ conduct.
What we do instead
Every customer gets invited to review publicly. No satisfaction pre-screen, no threshold, no diversion. That is the whole change, and it is the version that collects more reviews.
We keep the private feedback line — it is genuinely useful to hear a complaint the same hour it happens. It simply runs alongside the review invitation rather than in front of it. Nobody is filtered out of the public ask because of what they told you first.
Then we do the thing the evidence actually supports: reply to everything. Responding to reviews is worth +0.12 stars and 12% more reviews — nearly double the star gain filtering gives you, volume-positive instead of volume-negative, and the practice Google explicitly asks for.
Every claim on this page, with its source
We publish these because most content in this category does not, and because you should be able to check us.
- 01
Filtering costs 39% of review volume for 0.07 of a star
GatherUp measured ~10,000 locations across the 12 months before and after it removed sentiment gating from its own platform. Reviews went from 32,689 to 53,790. Average rating went from 4.66 to 4.59.
GatherUp, October 2019 · independently reported by Search Engine Land
- 02
Purchase likelihood peaks between 4.2 and 4.7 stars, then falls
Analysis across 40+ product categories found conversion rises with rating up to about 4.7 and then declines toward 5.0, because a perfect score reads as fabricated. 82% of shoppers deliberately seek out the negative reviews.
Spiegel Research Center, Northwestern University
- 03
Google's policy names the practice directly
Under Rating Manipulation, Google's prohibited content policy states merchants may not “discourage or prohibit negative reviews, or selectively solicit positive reviews from customers.” That last clause describes a sentiment gate exactly.
Google Maps prohibited & restricted content policy
- 04
An Australian company was fined $3 million for withholding invitations
Meriton altered or withheld guest email addresses so guests it expected to be unhappy never received the TripAdvisor review invitation. No reviews were deleted. The Federal Court held it misleading or deceptive conduct.
ACCC v Meriton Property Services, penalty July 2018
- 05
Suppressing negative feedback drew a $2.9 million penalty
HealthEngine did not publish around 17,000 reviews and edited roughly 3,000 to remove negative content or embellish them.
ACCC v HealthEngine, August 2020
- 06
The ACCC says vendors can be liable too, not just their clients
Its 2023 internet sweep states that third parties operating these business models “will also be contravening the ACL, or will be considered by the ACCC to have aided or abetted, or be knowingly concerned in” their clients' contraventions.
ACCC, Online reviews and testimonials internet sweep, December 2023
- 07
US regulators describe the gated funnel and prescribe the fix
The FTC Endorsement Guides give a worked example of inviting only happy customers to review, call it potentially unfair or deceptive, and then state plainly that simply inviting all recent purchasers would not be.
16 CFR § 255.2(e), example 11
- 08
Replying to reviews outperforms filtering on both measures
A difference-in-differences study across TripAdvisor and Expedia found responding to reviews produced a 0.12-star increase and 12% more reviews. Nearly double the star gain of filtering, and volume-positive rather than volume-negative.
Proserpio & Zervas, Marketing Science, 2017
Two honest caveats. The GatherUp dataset is the only published gated-versus-ungated comparison that exists, it comes from a vendor measuring its own platform, and it has never been replicated — it is the best available evidence and also the only available evidence. And applying the Meriton reasoning to a software product is an argument about where the law points, not a decided case. We would rather say that than overstate it.