Most advice here is about volume. The 2026 data says volume is the wrong target.
74% of consumers want to see reviews from the last three months. And 97% read reviews for local businesses at all.
Which produces a counterintuitive conclusion: a contractor with 40 reviews including six from this quarter frequently outperforms one with 200 where the newest is eight months old. The second profile reads as a business that changed hands or stopped caring, both to humans and to the ranking systems.
So this is not a campaign. It is a habit, and the variable that matters most is when you ask.
The short answer
Text a review link on the day of every completed job, asking one specific question, and respond to every review within a couple of days. That single habit, sustained, outperforms every campaign and tool. The businesses that struggle with reviews are almost never bad at asking; they are inconsistent about it.
Timing is most of it
The window is narrower than people think, and it is emotional rather than logistical.
The moment to ask is when the heat came back on. The customer is relieved, they like you, and the problem they had this morning is solved. That feeling decays fast. A week later the furnace working is simply normal again, and you are asking a stranger for a favour.
Same day, within a couple of hours of finishing. Not the next morning, not when you invoice.
In person first, then the link. "If you're happy with how that went, would you mind leaving us a quick Google review? I'll text you the link now." Asking face to face and then sending the link converts far better than a cold text, because the ask is already agreed.
One follow-up, two or three days later, to anyone who has not left one. Once. Then stop.
By the numbers
Test the timing yourself before believing anyone about it. Ask same-day for one month and at invoice time for another, and compare response rates. Most contractors find the difference is large enough to end the debate, and it is the cheapest experiment on this page.
Ask a question, not for a review
This is the part almost nobody optimises, and it now matters more than it used to.
"Please leave us a review" produces "Great service, highly recommend." Five stars, and close to worthless.
A specific question produces specific text. "Would you mind mentioning what the problem was and how quickly we got to you?" produces something like: "Furnace died Saturday night, Sam had it running by Sunday lunchtime. Fair price, tidy work."
That second review does four things the first does not:
It reassures a real prospect who has the same problem right now. It names the service, contributing to relevance for that query. It names the timing, which is what people actually want to know about a contractor. It can be quoted by an AI system recommending contractors, and generative AI tools jumped from 6% to 45% as a discovery source.
That last point is the newest and most underrated. Your review text is no longer only read by humans scrolling. It is being consumed by systems deciding who to recommend, and specific, detailed reviews are the ones that get used.
Rotate the question so your review profile does not read identically. What the problem was. How quickly you responded. What surprised them. Whether the price matched the quote.
The mechanics
Text beats email, comfortably. Higher open rates, less friction, and it arrives on the device they will use to write the review.
Send from the number they recognise. A text from an unknown number after a service visit reads as spam.
Use a direct link, shortened. Google provides a review link for your profile; anything that adds a step loses people. Do not send them to your homepage to find it.
Keep the message short.
Hi Mrs Chen, Sam from Northside. Glad we got the heat back on. If you have a minute, would you mind leaving a quick review? Even a line about how fast we got there helps a lot. [link] Thanks.
Supplement, do not rely on: QR codes on invoices, vehicle stickers, and a link in your email signature. These catch people you missed rather than replacing the ask.
Assign an owner. The technician who did the job asks in person; whoever does admin sends the link and the single follow-up. Unassigned, this stops in the first busy week.
What not to do
Three practices that are common, effective in the short term, and genuinely risky.
Do not offer incentives. Discounts, entries into a draw, free filters. Incentivised reviews violate Google's policies and enforcement can remove reviews or the listing. The volume gained is not worth the exposure, and the appeal process if it goes wrong is covered in Google Business Profile suspended.
Do not gate. Sending customers to a satisfaction survey first and only routing happy ones to Google is specifically prohibited. You may ask everyone, and you may resolve a complaint before asking, and you may not filter who gets the link based on how they feel.
Do not buy reviews or ask staff and family. Detection has improved, patterns are visible, and the damage from a purge exceeds the benefit.
Do not batch. Forty requests in one week produces a spike that looks engineered to Google and reads as suspicious to a customer scrolling ten reviews all dated the same Tuesday. Ten a month, every month, is the target shape.
Reviews bring people to your listing. Whether they become customers depends on what they find next. Our free check tests whether a visitor can enquire without phoning, whether hours and service areas are stated, and where the enquiry goes when it arrives. Twenty checks, about fifteen seconds.
Responding, including to the bad ones
Responses are a visible activity signal on a profile where activity now feeds rankings, and they are read by prospective customers more carefully than the reviews themselves.
Respond to everything, including five-star reviews. Two sentences. Name the job if you can, because it adds detail a stranger can use.
Respond within a couple of days. A profile where the owner replies promptly reads as a live business.
Do not use the same template. Ten identical "Thank you for your feedback!" replies look automated and undo the point.
On negative reviews, the response is not for the reviewer. It is for the next hundred people who read it, and what they are looking for is how you handle a problem. That is a big enough topic to have its own treatment in how to respond to bad reviews.
When you are starting from zero
A new profile with no reviews is the hardest position, and there is a legitimate way through it.
Start with recent customers. Anyone you served in the last month is fair to ask, and they still remember. Do not go back a year.
Space them out. Five in the first week on a brand-new profile looks exactly like what it looks like. Two or three a week is unremarkable.
Ask in person wherever possible while you are building the first ten. Conversion on a face-to-face ask followed by a link is far higher than a cold text.
Accept that the first ten take longest. Response rates improve once a profile has reviews, because leaving one feels normal rather than exposed.
Getting technicians to actually ask
The system fails at one predictable point: the person standing in the customer's house does not ask. Everything upstream of that is easy and everything downstream depends on it.
Three reasons technicians skip it, each with a different fix.
It feels like begging. The most common, and it is a framing problem. Reviews are how a good contractor gets found by the next person who needs one, and technicians generally accept that when it is put in those terms rather than as a marketing target.
They forget. Reliable and unglamorous. The fix is a physical prompt: a line on the job sheet, a step in the app before marking a job complete, or a card left with the invoice.
They are not sure the customer was happy. Legitimate, and the correct response is to ask anyway rather than to guess. Asking a mildly dissatisfied customer surfaces a complaint you can fix, which is more valuable than the review would have been. Quietly skipping them is how you end up with a problem you never heard about until it appeared in writing.
What works better than instruction: make it visible and make it shared. A number on a whiteboard for reviews received this month, per technician if the team is small enough. Not a bonus, which edges toward incentivised reviews and creates pressure to game it, just visibility. People do more of what gets noticed.
And handle the objection honestly. If a technician says customers find it awkward, ask them to try the exact wording above for two weeks and report back. Most discover it is considerably less awkward than they expected, because the customer has just had their problem solved and is well disposed toward the person who solved it.
What a review profile should look like
A useful target to aim at, rather than a number.
Volume relative to your market. Look at the three competitors ranking above you in the map pack at the centre of your service area. Match their count over time; you do not need to beat the largest business in the city, you need to be credible against the ones you actually compete with.
Newest review under three weeks old, always. This is the single most important characteristic and the one that requires ongoing effort rather than a campaign.
A distribution that looks human. Reviews spread evenly across months rather than clustered. Some four-star ones. A couple of critical ones handled well, which paradoxically increases trust because a profile of nothing but five stars reads as curated.
Text with specifics. Service names, timescales, neighbourhoods. This is what the question-based ask produces and what generic asks do not.
An owner response on every one. Including the positive ones, briefly, and not templated.
Photos from customers, which accumulate alongside reviews and often display prominently. Worth asking for occasionally: "if you got a photo of the new unit, feel free to add it."
A profile with those six characteristics outperforms a higher-rated, higher-count profile that is stale, generic and unanswered. That is a genuinely different target from "get more reviews," and it is achievable for a small contractor in a way that out-volumeing a large competitor is not.
Where else reviews do work
Worth knowing, because it changes how much the habit is worth relative to other marketing activity.
Local Services Ads. Your Google reviews influence which businesses are shown in that paid channel, so the same reviews pay into both organic and paid local visibility. The economics of that channel are in Google Local Services Ads for contractors.
Quote conversion. A customer deciding between two quotes checks reviews, and recent detailed ones addressing the specific job type do more than a star average. This is one of the reasons quotes go quiet, covered in why customers ghost after a quote.
AI recommendation. The shift from 6% to 45% for generative tools as a discovery source means review text is being read by systems answering "who should I call." Specific reviews are quotable; generic ones are not.
Your website. Reviews displayed on service pages address the same doubt at the point of decision.
Recruitment. Underrated. Technicians research employers, and a business with recent positive reviews describing good work is easier to hire into.
One habit, five returns, none of which require additional spend. That is a rare profile for a marketing activity and it is the honest reason this deserves protecting when the week gets busy.
The one-page system
Everything above, compressed to what actually goes on the wall:
- Technician asks in person before leaving, on every completed job.
- Link texted the same day, from the recognised number, with one specific question.
- One follow-up after two or three days to non-responders. Then stop.
- Owner responds to every review within two days, individually.
- Weekly check of newest review age. If over three weeks, the habit has slipped.
- Monthly review of requests sent versus reviews received.
Six steps, roughly fifteen minutes a week of admin, and it is the highest-return recurring habit available to a local trade. It improves ranking, conversion and AI recommendation simultaneously, and none of the three requires any spend.
The failure mode is always the same and it is worth naming so you can watch for it: this stops during busy season. The exact months when you are completing the most jobs, and therefore have the most opportunities to ask, are the months when nobody has time to ask. That is why it needs an owner and a prompt rather than good intentions, and why the newest-review-age check is the metric worth watching above all the others.
Measuring it
Four numbers, monthly, and they take five minutes:
Requests sent. The input you control. Reviews received. The output. Conversion rate. Below 20% suggests the timing or the message is wrong, not the customers. Newest review age. The single most important number on the list, given the recency finding, and the one nobody tracks.
That last one is worth putting somewhere visible. If your newest review is more than three or four weeks old, the habit has slipped, and it slips silently during exactly the busy periods when you are doing the most jobs worth reviewing.
Why this compounds
Reviews are unusual among marketing activities in that they improve three separate things at once, permanently.
Ranking. Prominence is largely reputation, and recency and response are both signals. The mechanics are in how to rank in the Google map pack.
Conversion. A prospect comparing three contractors with 4.8, 4.9 and 4.7 stars reads the recent text, not the average. Detail decides it.
AI recommendation. The 6% to 45% shift means your review text is being read by systems answering "who should I call for a furnace repair." Specific reviews are quotable; generic ones are not.
No other fifteen-minute weekly habit affects all three. Which is why, of everything in a local marketing programme, the review habit is the one worth protecting when things get busy, and it is reliably the first thing dropped.
The wider profile work it feeds into is in Google Business Profile for contractors, and if the enquiries those reviews generate are not being answered quickly, that is the prior problem and it is quantified in how many leads is my business losing.