Why you cannot ask only your happy customers

This practice has a name

Deciding whether to ask for a review based on how happy the customer seems is called review gating. Google’s prohibited and restricted content policy says it plainly: you may not solicit reviews only from customers likely to leave a positive one, and you may not put any filtering step in front of who gets sent to Google.

The usual first reaction is “how is that a violation? I am not faking anything.” And you are not — every review is a real person describing a real visit. The problem is not the reviews, it is the sample. If you systematically let only some people speak, the average on your listing stops representing your customers and starts representing your filter. That is the thing Google prohibits.

What it looks like — including the version you may be doing

The obvious version is a toggle in a piece of software, but far more often it hides inside everyday behaviour. All of these are gating:

• Asking “happy with everything today?” first, and only handing over the code if the answer is yes. The most common version, the one that feels least like a violation, and the most textbook example there is.
• A star-rating splitter page. The customer taps a star on your page; four and up go to Google, three and below go to a private “tell us more” form. This is gating in its purest form.
• Only inviting regulars, members or people you know. Swapping “happy” for “friendly with us” as the filter changes nothing.
• Staff reading the room and deciding whether to mention it. No system at all, identical outcome — and hard to claim you did not know.
• Sounding them out privately first and dropping it if the answer is lukewarm. Testing the water is still filtering.
• Only handing over the code after a customer has praised you. This one reads as good timing, but it ties the decision to ask to the level of satisfaction.

Look at the last three: no software, no toggle, purely human behaviour. Gating is a practice, not a feature — you can do it perfectly well with no tool at all.

How Google finds it, and what it costs

Google does not need to see what happens in your shop. The data says it on its own.

1. The distribution goes wrong. A normal business has a J-shaped curve: five stars most common, one star second, a dip in the middle. A listing with nothing but fours and fives and almost no low scores has a shape that is not natural.
2. Splitter pages get found. A star-rating splitter is a publicly reachable web page and gets crawled. Businesses using third-party gating software usually get caught in batches, when the software itself is identified.
3. Customers report it. A decent share of the people diverted into a private form go straight to Google anyway and say in the review that you would not let them post — and that review is itself the evidence.
4. Staff and ex-staff talk. This kind of practice usually lives in a training script, and scripts leave a trail.

What follows is graduated: at the mild end, affected reviews are removed in bulk (genuine ones alongside them); harder, the listing gets flagged and new reviews restricted; hardest, the profile is suspended. The cost is never one review. It is the whole page. What to do if it is suspended is in Google Business Profile suspended? What to do; how to work out why reviews vanished in bulk is in Your Google reviews disappeared — what happened.

Is filtering even worth it? Here is the maths

Set compliance aside for a moment and do it as pure business arithmetic. Say fifty customers a day, forty-five happy and five unhappy, and three in ten of those you ask actually post.

Look at the second line: the 5.0 that filtering buys you converts worse than a 4.7. Consumer research keeps landing on the same finding — a perfect score makes people suspicious, and somewhere between 4.5 and 4.8 is where people actually believe you. Which means gating risks a suspension in exchange for a rating that performs worse. It does not pay, quite apart from whether it is allowed.

What you are allowed to do

What Google bans is filtering by satisfaction, not asking. All of the following are entirely compliant, and they work better anyway:

• Same sentence, same code, every customer. Treating everyone alike is the one hard requirement.
• Choose the moment, not the person. Ask at the point where the thing is done — keys handed back, main course halfway through, follow-up visit confirming a good recovery. A moment applies to everybody; a mood does not.
• Collect positive and negative on the same page. Let the customer choose, rather than deciding for them whether they get to post.
• Keep internal satisfaction surveys completely separate from public reviews. Run NPS, send questionnaires — just never use the result to decide who gets sent to Google.
• Ask the unhappy ones too. Counter-intuitive, but a four-star saying “there was a problem and the owner fixed it on the spot” often persuades better than your tenth five-star.

This is how Review Helper is built, too: the top of the page holds the positive options, and a collapsed section under a neutral heading holds ten negative ones, each mutually exclusive with its positive twin. Every customer gets the same page and the same options, and what comes out is a draft they can rewrite — posting it is their call. That is not product timidity. It is the line you cannot cross.

One compliant sentence

Swap “pick the person” for “pick the moment” and the sentence actually gets easier, because you no longer have to read anyone’s mood first:

“That is everything sorted. If you have thirty seconds, would you scan this and say how it went — good or bad, we read all of it.”

The words “good or bad” are doing the work. They settle the compliance question in one breath, and they lower the social cost for the customer: being asked to say something nice is awkward, being asked to say something honest is not. When to ask, how to handle a no, and how to get staff comfortable saying it are in How to ask customers for Google Reviews.

A checklist

• Is there any step where you judge the mood first and then decide whether to ask? That is gating.
• Does every customer get the same code, the same sentence, the same page?
• If a customer wants to say something negative, is there a route that does not pass through you?
• Are internal satisfaction scores being used to decide who gets sent to Google?
• If you use third-party review software, does it have a “send four stars and up to Google” feature? Turn it off.
• Does your staff script say “ask everyone” or “ask when it feels right”?

One line to remember the boundary by: you may choose when to ask; you may not choose whom to ask. Hold that, and the rest is turning the ask into a habit — the pacing for a brand new listing is in From 0 to 50 Google reviews in 90 days, where the discount-for-a-review line sits is in Can you offer a discount for a Google review?, and what to do when a bad one does arrive is in You got a one-star note. Now what?