Blog · 6 October 2026 · 9 min read
Review Velocity, and How to Build It Without Risk
Almost every business owner I speak to thinks about reviews as a total. They want to get to fifty, or to beat the competitor sitting on a hundred and twenty. That number matters, but it is not the number that moves rankings fastest, and chasing it the obvious way is how profiles get filtered.
The thing that actually behaves like a ranking signal is velocity: how regularly new reviews arrive, over how long, from accounts that look like real people who were really there. Here is how that works in practice, and how to build it without taking on risk.
Why a burst of reviews does less than a trickle
Picture two competing clinics. One collects thirty reviews over a single fortnight after the owner texted every patient from the last two years. The other collects three or four a month, every month, for a year, finishing with roughly forty.
The second profile will almost always out-perform the first, and often by a lot. Three things are happening. The steady profile looks like a business with continuous custom, which is what Google is trying to identify. Recency carries weight, and a listing whose newest review is eleven months old reads as a business that may have stopped trading. And the burst pattern is a well-known manipulation signature, so a share of those thirty often get filtered out within weeks, sometimes without any notification.
A real example. A home services client arrived with ninety-one reviews and no ranking to show for it. The history was a wall of reviews across two weeks in the previous year, then nothing. We stopped chasing volume entirely and installed a process that produced four or five a month. Six months later the profile had about a hundred and twenty, barely a third more, and it had moved from outside the pack to second for its main term across most of the service area. The total was not what changed. The pattern was.
This is also why I will not take on a client who wants reviews bought. It is the fastest route to a suspension, it is trivially detectable, and it undoes the rest of the work. The honest version of how I build local rankings for service businesses is slower in month one and considerably faster by month six.
The ask: timing beats wording
Most advice obsesses over the script. In practice when you ask matters more than how.
There is a short window after a job completes when the customer is at peak satisfaction, and it closes fast. For an emergency call-out, the window is minutes: the drain is clear, the relief is real, and the technician is still standing there. For a dental implant it might be a week later, after the follow-up. For a house cleaning, it is the moment the client walks back in.
So the process has to be owned by whoever is present at that moment, not by an office that batches requests on a Friday. The highest-performing setup I have put in place was for a workshop: the technician handed over the keys with a small card carrying a QR code straight to the review form, and said one line, that reviews are how people find them. No app, no automation, no incentive. It went from about one review a month to six, and it never dipped afterwards because it was part of handing over the keys rather than a campaign somebody had to remember.
Two things to avoid, firmly. Do not offer anything in exchange, not a discount, not entry into a draw, not a free coffee. It violates policy, and reviews traceable to an incentive get removed along with, sometimes, the ability to receive new ones. And do not screen customers by asking how happy they are first and only sending the link to the happy ones. That practice, review gating, is explicitly against policy and is easy to spot in the ratings distribution.
What a useful review actually contains
A five-star review saying great service is worth less than a four-star review saying the engineer arrived the same afternoon in Johar Town, found the fault in the condenser and had it running before evening.
The second one contains the service, the area, the problem and the outcome. That is text Google can match against queries, and it is text that AI answers can draw on when summarising who to call. The first one is a number with no information in it.
You cannot script this without crossing into writing reviews for people, which you must not do. What you can do is prompt memory. Instead of asking for a review, ask the customer whether they would mind mentioning what the problem was and how quickly it got sorted. People then write naturally and specifically, because you have reminded them what happened rather than told them what to say.
A pest control client in Birmingham did exactly this with one changed sentence in their follow-up message. The review count did not change much. The content did: within three months most new reviews named a suburb and a pest type, and the profile started surfacing for searches combining the two.
Replies are not customer service, they are content
Owner responses sit on the profile permanently, they are indexed, and almost everybody wastes them. Thank you for your feedback, repeated eighty times, is a wasted field.
A good reply does three things in two or three sentences. It thanks the person naturally. It restates the job in plain language, which puts the service and often the area into text you control. And it points at the next step for whoever is reading it later, because the real audience for a reply is not the reviewer, it is the next prospect comparing you against two others.
Reply to negative reviews too, and reply calmly. Prospects read the one-star reviews first, and what they are judging is not the complaint, it is your response to it. A measured reply that acknowledges the issue and states what was done converts better than a profile with no negative reviews at all, which many people now read as suspicious. I have seen a one-star review with a decent reply do more good than harm.
Keep replies prompt but not instant across the board. Answering a batch of thirty old reviews in one sitting is its own odd pattern. Work through them over a fortnight.
How to measure whether any of it is working
Track three things monthly, and ignore the rest. New reviews received, because that is the velocity. The proportion mentioning a specific service or area, because that is the quality. And your map pack position across the service area, because that is the outcome.
The lag is real: expect two to three months before a sustained change in velocity reads clearly in position, longer in a competitive category. If you are measuring position from a single point you will not see most of it, which is why I measure across a grid of points instead of one.
If velocity is up, content is improving and position has not moved at all after three months, reviews were not your constraint. That is a genuinely useful finding, and it usually points at categories, proximity or a profile with half its fields empty, the same things covered in a full Google Business Profile optimisation.
This is also the part of the work that cannot be bought, rushed or outsourced to a tool, which is precisely why it keeps working. If you want the process built around how your business actually finishes a job, that is the kind of thing I set up for clients from Karachi to Kansas City, and it is covered in the way I work with service businesses.
Written by Waseem Abbas. More about how I work, or see client results with the screenshots.
WhatsApp me and I will look at your review pattern
Tell me your niche and your market and I will tell you which of these are worth registering for you, in what order, and which to leave alone.