What Can You Ask Customers to Say in a Google Review?

Ask every customer for an honest Google review and give them a direct link to the right branch. Google's Maps user-generated content policy, re-read November 1, 2026, allows review requests only without "attempting to influence the rating or the contents of the review." In the positive home-service reviews we studied, 47.6% already named a person. Towns almost never came up, so they belong on pages you write.

Last verified: November 1, 2026.

Start this week. Pull every review request your company sends, from the tech's script and the text template to the printed card and the page a QR code opens, and strike any line that asks the customer to put something in the review, whether that's a name, the town, a service, how fast you were or a star rating. Replace it with one open line: "If you'd like to share an honest review of today's service, you can tap here. It takes you straight to our Google page, and you can always do it later."

What Google's policy says about review content

Google keeps its rules on asking for reviews in the Rating manipulation section of its Maps user-generated content policy, and one of them settles the question. Merchants shouldn't "request that specific content be included," and the example Google gives is content "that identifies a staff member." The contents clause quoted above covers the rest without any interpretation from us, because asking a customer to mention the town, the service or the stars is an attempt to influence what the review says.

The other rules cover how you ask. Merchants can't have staff solicit "a certain number of reviews," so no tech gets a review quota. Asking at the door is fine, but Google says merchants shouldn't require or pressure a review "while on the premises," which is why the open line tells the customer they can do it later. Nor can you ask only the customers you expect to be happy: businesses may not "selectively solicit positive reviews from customers," so the job that went sideways gets the same ask as the one that went perfectly.

Customer incentives in exchange for a review are prohibited. Google's own Tips to get more reviews page calls them "strictly prohibited" and tells businesses to ask customers to "visit a Google link or scan a QR code."

Search Engine Roundtable first reported the staff clauses on April 17, 2026; Google's page carries no date of its own and doesn't define "specific content." Our recommendation is straightforward: an ask that points the customer at a topic, even a harmless one like how quickly you showed up, counts as specific, while an ask for an honest review of the visit doesn't.

What to say when you ask

Every template below asks for an honest review and names no topic and no star rating.

• In person: the open line above. On a printed QR card, swap "tap" for "scan."
• Text: "Thanks for having us out today. If you'd like to share an honest review of the visit, here's our Google page: [link]"
• Email, with the subject line "Thanks for having us out": "Thanks for choosing [Company] today. If you'd like to share an honest review of the visit, this link opens our Google page: [link]. Write it whenever it suits you."

Two details matter more than the wording. The link has to open that location's own Business Profile, not the brand's main listing or a branch two towns over, and every customer on every finished job gets the same words. Our point-of-service review playbook covers timing, the handoff and who makes the ask.

Almost half of the positive reviews we studied already name someone

Negative reviews name someone far less often than that: 12.6% of one- and two-star reviews did, so a name mostly comes with praise. Nor is the positive rate one company's habit: it barely moves without the largest company, which supplies about a third of those reviews, and it held through the year.

An outside study points the same way. Reviewflowz and Cartelis found a person named in 28% of positive Google reviews across the US and Europe, rising to 66% in US home services, PPC Land reported on October 28. Their method isn't ours, so the numbers don't line up, but both studies say customers naming the person who did the job is normal in home services.

The study also found that reviews naming a person disappeared at about the same rate as reviews that didn't once business, age and length were held constant (1.01 times). Customers naturally recognize good service by naming the person who delivered it. Let them write in their own words and handle employee attribution through the review link or job record.

The town and the specialty services rarely show up in reviews

Start with the town. Even counting the profile's city or any place on its service-area list, which runs to about 12 places per profile, only 0.79% of positive reviews named a town. The profile's own city alone turned up in about 1 in 200 text reviews (0.49%).

The specialty work owners list on their profiles to win bigger jobs barely shows up either. Heat pumps are listed on 62.5% of HVAC profiles and mentioned in 0.5% of HVAC reviews, and mini-splits or ductless systems are on 75.0% of the profiles and in 0.3% of reviews.

The rest of the chart runs the same way. AC and furnaces are each listed on 75.0% of the profiles, against 11.0% and 8.9% of reviews; maintenance or tune-up runs 81.3% against 5.8%, ductwork 75.0% against 2.8%, thermostats 78.1% against 1.4%, and air quality 46.9% against 0.8%.

Plumbing profiles show the same gap. Drain, leak, water heater and toilet work are each listed on 92.3% of the profiles, against 6.7%, 6.4%, 5.4% and 4.6% of text reviews. Sewer runs 90.4% against 2.3%, softener or filtration 82.7% against 1.4%, tankless 65.4% against 0.9%, camera inspection 48.1% against 1.0%, and hydro jetting, listed on 73.1% of the profiles, turns up in just 0.3% of reviews.

Complaints describe the job far more often than praise does. One-star reviews ran a median 96 words, and 72.2% of them used a job, equipment or repair word; five-star reviews ran a median 22 words, and 36.8% did. When a one- or two-star review comes in, treat it as a job record: reply, fix the job, and check whether one of your own pages set the wrong expectation.

Put the town and the job on pages you write

You can't ask for the town or the service in a review, so put them where you choose every word: the website and the Business Profile. On each location page, name the towns that branch serves and show recent jobs there. Every job you want more of, heat pump installs or hydro jetting, gets its own service page with real recent work and photos, because customers aren't going to write that page for you. Then make the profile's services list match those pages, described in your own words.

In our own research panel, ChatGPT started linking business sites far more often in October. Its local answers in the Cheers Market Baseline switched to a per-business layout on October 9, 2026, and the share of US ChatGPT answers that linked the website of at least one business on the answer's own map cards rose from 41.7% to 78.1%.

Engines pick which page to cite, but you decide which pages exist for them to pick. In September, service, location and service-plus-city pages drew 48% of AI citations to contractor websites across four AI engines, against 43.1% for homepages and 0.7% for testimonial pages on contractors' own sites. That's why the towns you serve belong on the location page, not in a customer's review.

Reviews are a central part of that reputation. Google AI Mode showed a Google rating and review count for 90% of the businesses it named in our September sample, and how Google reviews affect AI recommendations has the rest of those numbers. Reviews carry the rating and the count AI shows next to your name; your pages have to carry the towns and the jobs.

How to credit a tech when the review doesn't name anyone

Cheers can credit the employee through the link they shared or the job record, including when the customer leaves their name out of the review. That gives a manager a useful record for recognition and coaching while the customer writes freely. Tracking the result does not require a review quota. Our employee recognition and review guide explains how we apply that distinction to bonuses and leaderboards.

If you want AI to know who does the work, put the tech on a page you control. A bio lets you describe the person doing the job without asking a customer to supply the wording, and when technician bios help covers what one should carry.

A review-request audit you can run this week

• Collect every script and template your company uses, including printed cards and the page a QR code opens.
• Strike any line that tells the customer what to write about or what to rate. "Mention my name," "say we serve [town]," "talk about the new heat pump" and "five stars would really help" all go.
• Open each review link on a phone and confirm it lands on the right location's profile.
• Confirm every customer on every finished job gets asked, including the jobs that went badly.
• Count how many of your last 100 positive reviews name someone, per brand or branch. In our sample the middle half of the 13 companies with 100 or more positive reviews landed between about 46% and 51%. Treat that as context, never as a target for a tech, and don't answer a low number with a name prompt.

Running several brands? Write one review-request standard for the portfolio and check each acquired brand's templates against it during onboarding, since scripts tend to survive acquisitions untouched.

Last, read what customers already praise in recent reviews and carry those themes onto your own pages. Turning reviews into AI search content shows how to do that without scripting anyone.

Method

These are Google reviews with text on home-service Business Profiles connected to Cheers, in the US and Canada, posted October 1, 2025 to September 30, 2026; more than nine in ten come from US locations. We left out deleted and replaced reviews and queried the data on November 1, 2026. The profiles belong mostly to established companies, so read the rates as describing these profiles, not every home-service business, and as what customers wrote, not how they were asked.

The named rate covers 22,083 four- and five-star reviews at 113 locations of 19 companies, and the one- and two-star rate covers 910 reviews at 76 locations of 17 companies; across every star band, 46.0% of 23,147 text reviews named someone. Without the largest company, which supplies about a third of the positive reviews, the named rate is 47.5%. The rate was 47.4% for the 11,677 positive reviews posted before April 17, 2026, and 47.8% for the 10,406 posted on or after that date.

Names come from automated name extraction, which can miss a nickname or count a customer's own family member, so "names a person" isn't strictly "names the tech." A review counts as naming a town when it contains the profile's city or a place on its service-area list. A town that's also a first name can add a false hit, and the match misses neighborhoods that aren't on the list. The stricter own-city check covered 23,089 text reviews at 104 locations with a usable city name (19 companies).

Service and job-word mentions are keyword matches, so they count negations ("didn't need a new heat pump"), and "duct" also matches "ductless." The job-word rates rest on 747 one-star reviews at 71 locations and 21,645 five-star reviews at 113. We dropped any service row with mentions at fewer than 10 locations. HVAC rests on 16,067 text reviews on 32 profiles at 7 companies, and plumbing on 6,204 reviews on 52 profiles at 8. Two companies' profiles hold most of the HVAC reviews. Without the largest HVAC company, heat pumps turned up in 0.5% of HVAC reviews and mini-splits in 0.2%. Counting only profiles that list the service, the rates were 0.5% for heat pumps, 0.3% for mini-splits, 0.4% for hydro jetting and 0.8% for tankless.

AI answer figures come from the Cheers Market Baseline, which every week asks ChatGPT, Google AI Mode and Gemini the same 32 home-service questions in 100 metro areas, 60 in the US and 40 in Canada. The October link shares compare 1,910 US ChatGPT answers from October 2 to 8 with 6,307 from October 9 to 31. "Linked a business's website" means a cited domain matched the site of a business on that answer's own map cards, so it's a floor, and it doesn't show which page got linked. Directory links grew faster, so across all 100 metros the share of ChatGPT's links going to a carded business's own site fell from 28.7% to 18.1%. Page types come from our published September study (74,969 citations in 34,742 answers from ChatGPT, Gemini, Perplexity and Google AI Mode in US and Canadian metros, August 31 to September 20, 2026), and the AI Mode rating figure from 9,435 AI Mode answers in the same window. These AI numbers are saved or already published aggregates; we didn't re-run them for this article.

Reviewflowz and Cartelis looked at 4,876,933 Google reviews on 9,048 profiles, crawled August 11 to 24, 2026, and published the study October 16; we cite PPC Land's report because we couldn't find the study's own page. Their 1.01 deletion ratio holds the business, the review's age and its length constant (95% interval 0.91 to 1.11). In raw counts it was 1.32 times, which the report puts down to named reviews sitting in businesses and periods with higher deletion rates overall.

Sources

• Maps user-contributed content policy, Rating manipulation section, Google, undated. The rules on soliciting reviews, including no requested content and no staff quotas.
• Tips to get more reviews, Google Business Profile Help, undated. The allowed Google link or QR code ask and the incentive ban.
• Google Updates Reviews Policy To Restrict Staff Mentions & Solicitations, Search Engine Roundtable, April 17, 2026. When the staff clauses were first reported.
• Google Maps businesses lose 1.95% of new reviews in 7 days, study finds, PPC Land, October 28, 2026. The Reviewflowz and Cartelis naming and deletion figures.

Sources read November 1, 2026.

Dylan Allen-Arnegård is the CEO & Co-Founder of Cheers, the done-for-you local visibility partner for established home-service companies and PE-backed portfolios.