On July 3, 2026, a business owner posted in the Google Business Profile forums that their review count had gone from roughly 4,651 to 63. Inside 24 hours. No suspension notice, no warning email, nothing to appeal.
They weren’t alone. Reports piled up across the forums that week, and Search Engine Roundtable documented the wave as Google confirmed it was looking into it. Google’s line was that when its systems flag suspicious activity, it can strip reviews and pause new ones on a profile while it sorts things out. Most of those reviews came back within a week or so. Plenty of owners spent that week refreshing a page and feeling sick.
Here’s the thing that episode exposed. If a week of downtime on your review profile makes your phone stop ringing, you don’t have a trust asset. You have a rental agreement with a landlord who never answers the door.
4,651 Reviews Gone in a Day
Most local marketing advice treats your star rating as the finish line. Hit 4.7, collect the leads, done.
But your rating lives on someone else’s server, gets filtered by someone else’s spam model, and can be wiped by a competitor’s sabotage campaign or an overcorrecting algorithm on a random Tuesday in July. Google has leaned harder on automated filtering through 2026, partly because regulators in the US and Europe have been pushing platforms to clean up fake reviews. Aggressive filters catch fakes. They also catch you.
So here’s a better question to run your marketing against: if my Google profile vanished tomorrow, what would still make a stranger pick me? The answer to that is your actual trust equity, and reviews are one deposit into the account rather than the account itself. Building that equity should be a core part of any sustainable business growth plan.
Your Customers Stopped Looking Where You Think They Look
BrightLocal’s Local Consumer Review Survey 2026, run across a representative panel of 1,002 US adults, turned up a shift that should reset how you budget your attention.
The share of consumers using AI tools like ChatGPT and Google’s AI Mode to find local businesses went from 6% to 45% in a single year. Over the same stretch, Google’s share of local business discovery slid from 83% to 71%. The average consumer now touches around six different review sources before deciding.
| Discovery channel | 2025 | 2026 |
| 83% | 71% | |
| AI tools (ChatGPT, AI Mode, Gemini) | 6% | 45% |
Read that table as a warning about concentration, not as a reason to chase a new platform. Ninety-seven percent of consumers still read reviews before choosing a local business. What changed is where those reviews get summarized, quoted, and repackaged before a human ever sees your profile.
An AI assistant answering “who’s a good plumber in Coquitlam” is pulling from review text, your website, directory listings, and whatever else it can reach. Thin, generic review text gives it nothing to work with.
The Reviews That Actually Move People
Volume gets you past the filter a shopper applies in the first three seconds. What happens after that depends entirely on what your reviews actually say.
Specifics Beat Superlatives Every Time
Ask ten customers for a review, and eight will write some version of “great service, highly recommend.” Those are polite. They’re also worthless to the next reader, who already assumes you’re at least okay or you wouldn’t have 4.6 stars.
The reviews that do work name something. A staff member by first name. What the problem was. How long it took. What went wrong and how you handled it.
Pull up the profile of a busy independent shop, a maple ridge dentist, a family accountant, and the same pattern shows up. The reviews people quote back to you in the intake conversation are the ones that describe a specific situation the reader recognizes as their own. An emergency on a Saturday. A nervous kid. A billing question that got answered honestly.
You can influence this. Not by scripting anyone, which is both gross and against most platform policies, but by changing the ask. “Would you mind leaving us a review?” produces mush. “If you have a minute, it helps other people a lot if you mention what you came in for and how it went” produces something a stranger can use.
Same effort. Completely different output.
Old Reviews Read Like Old Milk
Consumers in the 2026 survey put more weight on recency than they used to, and the bar on star ratings keeps creeping up. A 4.8 average built entirely in 2023 reads worse to a careful shopper than a 4.6 with fifteen reviews from the last two months.
Recency is also the one metric that proves nothing bad happened recently. Your five-year-old reviews say your business was good under whoever was working then. Ten reviews from the last quarter say you’re good now, with the current staff, at the current volume.
So if your last review is from February, fix that before you spend another minute worrying about the decimal on your average.
Replying Works, and the Reason Is Kind of Uncomfortable
Davide Proserpio at USC and Georgios Zervas at Boston University ran what’s still the cleanest study on owner responses, published in Marketing Science and summarized by INFORMS. They analyzed tens of thousands of hotel reviews on TripAdvisor and tracked what happened when management started replying.
Responding hotels saw review volume rise about 12% and average ratings climb by roughly 0.12 stars.
The mechanism is the interesting part. Service quality had nothing to do with it. Unhappy customers carrying weak, unsubstantiated complaints simply became less likely to post at all once they could see an owner was reading and would answer in public. Attentiveness changed who bothered to write.
The researchers also found responding businesses got fewer negative reviews, but longer and more detailed ones when they came. That’s the trade. You lose the drive-by one-star and gain the serious complaint written by someone who knows you’ll read it.
I’d take that trade every time. A detailed complaint is free consulting. A vague one-star is just weather.
Reply to the good ones too, and keep it short. Nobody trusts a business that writes four paragraphs thanking someone for a cleaning appointment.
Where the Regulators Draw the Line
This part stopped being theoretical.
The FTC’s Rule on the Use of Consumer Reviews and Testimonials took effect on October 21, 2024. It bans buying or selling fake reviews, reviews from people who never used the service, insider reviews from staff or their relatives without disclosure, and incentives tied to leaving a positive review specifically. On December 22, 2025, the FTC sent its first round of warning letters under the rule to ten companies, giving them five days to confirm corrective action. Civil penalties run past $53,000 per violation.
Canada got there first in spirit. The Competition Bureau warned businesses in January 2024 that reviews posted by employees without disclosing the employment relationship can trigger enforcement under the Competition Act, and that owners who look the other way are on the hook. The Bureau’s earlier action against Montreal’s Amp Me Inc. ended with a $310,000 partial penalty plus $40,000 in costs, against an originally imposed $1.5 million.
The consumer side backs this up. BrightLocal found that 93% of people think someone should be policing fake reviews, and a large share put that responsibility on businesses themselves rather than on platforms or regulators.
Practical translation for a five-person shop: don’t let your office manager review you, don’t offer a discount for five stars specifically, and don’t buy anything from anyone who emails you offering reviews. A gift card for any honest review is generally fine. A gift card for a good review is not.
A 30-Day Plan That Costs You Nothing
| Week | Do this | Why it matters |
| 1 | Screenshot your current review counts on every platform and save the file | If reviews vanish, you need proof of what existed. Owners hit in July 2026 with no records had nothing to show support |
| 2 | Reply to every unanswered review going back six months, two sentences each | Signals attentiveness, which changes who bothers to complain publicly |
| 3 | Rewrite your review request to ask for the specific situation, not a rating | Produces review text a human and an AI summary can both use |
| 4 | Fix the three things your negative reviews keep naming | The complaint pattern is your actual product roadmap |
Nothing here needs a tool, an agency, or a budget line. Week 4 is the one people skip and the one that matters most.
The Part Nobody Can Take From You
Everything above sits on rented land. Here’s the part you own outright.
The businesses that came through the July outage without a dip were the ones where a decent share of new customers arrived already convinced, because a neighbor mentioned them, because their name shows up in local Facebook groups, or because someone remembered the front desk person’s name. Reviews confirmed a decision those customers had mostly already made.
That’s what a healthy funnel looks like, and the same trust-first thinking applies when building better workspace for small business.
Build for that, and your star rating becomes a nice-to-have instead of a life support system. Keep asking for reviews. Keep replying. Just stop treating a number on someone else’s platform as the thing your business rests on.
Be First to Comment