The terms, precisely
The distinctions here are narrow and they decide whether a practice is allowed.
- Review gating
- Screening customers by how satisfied they are and routing only the satisfied ones toward a public review, while the dissatisfied ones are directed to a private channel instead. The screening is the violation, not the private channel.
- Selective solicitation
- The broader form: asking only some customers for reviews, chosen on any basis correlated with how happy they are. This includes the entirely manual version where a receptionist asks the pleasant customers and not the difficult ones.
- Feedback collection
- Asking customers privately how things went. Entirely permitted, genuinely useful, and it becomes gating the moment the answer determines who receives the public review link.
- Review incentives
- Offering anything of value in exchange for a review — a discount, entry to a draw, a free coffee. A separate prohibition from gating and generally a clearer one, though the two are often sold together.
Also called: selective solicitation · sentiment filtering
The line, and where it actually falls
Google's review policies prohibit soliciting reviews selectively and discouraging or prohibiting negative reviews. That is the whole basis, and it is short enough that most disagreements come from wanting a different rule rather than from ambiguity in this one.
What is allowed: asking every customer for a review. Asking every customer for private feedback. Asking every customer for both. Choosing the moment, the channel and the wording. Following up once.
What is not: using a satisfaction answer to decide who sees the review link. Sending a one-to-five-star widget where four and five go to Google and one to three go to a form. Asking only the customers whose jobs went well.
The awkward part is the third example. A business that asks at the counter and skips the difficult customers is doing exactly the prohibited thing manually, and nobody thinks of it that way. The software version is more visible, more systematic and easier to detect, but it is not a different practice.
Allowed, not allowed, and the grey strip
The middle column is where most disputes actually live.
| Practice | Position | Why |
|---|---|---|
| Ask every customer for a public review | Allowed | No selection is happening |
| Ask every customer for private feedback too | Allowed | Both asks go to everyone |
| Star widget: 4–5 to Google, 1–3 to a form | Prohibited | The textbook implementation of gating |
| Ask only customers whose job went well | Prohibited | Selective solicitation, whether a person or software decides |
| Ask only customers of one service line | Grey | Fine if the basis is genuinely operational, not if the service line is a proxy for satisfaction |
| Delay the ask for customers with an open complaint | Grey | Defensible as timing; indefensible if the delay is permanent |
| Offer a discount for reviewing | Prohibited | A separate rule about incentives |
The two grey rows are grey because intent decides them, and intent is exactly what a policy cannot inspect. The practical test: would you be comfortable explaining the rule to the customer who did not get asked?
The arguments for gating, taken seriously
- One angry customer can drag a small average down for a year.
- True, and it is the strongest version of this argument. A business with forty reviews feels a single one-star far more than a chain with four thousand. The answer is not filtering — it is volume, because a steady stream of genuine reviews dilutes any single outlier faster than any gate protects against it. The businesses hurt most by one bad review are the ones with the fewest reviews, and gating keeps them there.
- Everyone does it.
- A meaningful share of the category does enable it, sometimes without naming it. That does not change the position, and it does mean the risk is unevenly enforced — which is a genuinely uncomfortable place for a competitor who does not offer it to stand.
- I only want to fix problems before they become public.
- Then do exactly that, and send the review request as well. Fixing the problem is good practice; withholding the link is the violation. These are separable and businesses conflate them constantly.
- What is the actual consequence?
- Enforcement is inconsistent and the range runs from reviews being removed to a profile being suspended. Anybody quoting you a probability is guessing. What is not a guess is that the risk sits with the business, not with the software vendor who enabled it.
What to do instead, in order of effect
Five things, all permitted, and the first two outperform gating anyway.
- Ask everyone, consistently, foreverVolume is the actual defence against a bad review. Two hundred reviews absorbs a one-star; thirty does not, and the difference is entirely whether the asking kept happening.
- Ask for private feedback at the same timeSame message, two links, no branching. You get the operational signal and the customer gets both routes, which is the compliant version of what gating was trying to achieve.
- Reply to the negative ones, publicly and brieflyA calm, specific reply does more for a reader than the review does against you. It is free, it is visible, and most businesses either ignore reviews or argue with them.
- Fix the thing the review is aboutObvious and rarely done. Recurring complaints in reviews are a product problem wearing a marketing costume.
- Get the timing right per serviceThe single largest permitted lever on response rate. It is worth more than any wording change and it is where most of the actual engineering in this category should go.
The first item is the whole argument. Gating protects an average; volume makes the average robust, and only one of those is allowed.
Questions
Is gating illegal, or just against policy?
What about other review platforms?
Can I ask customers to remove a review?
Does refusing to build this cost you customers?
The compliant version, in practice
Timing, channel, and wording that does not sound like a form letter.