Why a Steady Trickle of Reviews Beats a Big Batch
Most trades treat reviews as a job. You sit down one wet Tuesday, work through eighteen months of invoices, text everybody, and collect a dozen reviews in a fortnight. Then nothing for a year.
That pattern is worse than it looks, and not for the reason you have been told. The usual warning is that Google filters a batch as spam. The evidence for that is thinner than the people repeating it admit. The stronger argument is duller: reviews go stale, in the eyes of customers and of the algorithm, and a batch guarantees they all go stale on the same day.
The short version
- Two review-pattern factors sit in the top 15 local pack factors: recency at 11, steady growth at 14.
- 74% of US consumers look for reviews from the last three months, and the share wanting one from the last two weeks jumped from 20% to 32% in a year.
- Google's policy bans unusual review volumes that indicate manipulation. Volume alone is not the offence, and Google's missing-reviews page never lists batching as a cause.
- Review posting blocks are real, but the tracked cases involve months of abnormal volume, not one busy week.
- The evidence for velocity affecting rankings is expert opinion plus correlation. Nobody has proved causation.
Does review recency actually affect local rankings?
The people who study local search think so, and they think it twice. In the local pack factor ordering published by BrightLocal in January 2026, recency of reviews sits at number 11 and steady growth of reviews over time sits at number 14, as two distinct factors alongside review count and star rating.
That is the important detail. Four review dimensions appear in that top 15, and only two are about the reviews themselves. The other two are about the pattern: how recent, and how consistent.
Darren Shaw of Whitespark has argued this for longer than most. In May 2025 he called review recency the most underrated local ranking factor, noting it placed only 20th in the 2023 expert survey while he rates it inside his personal top five. His summary is blunt.
"The moment you stop getting new reviews, you're going to see your local rankings start to slip."
He goes further than the data does, and it is worth flagging where. Shaw also argues that a negative review beats no new review at all, because recency counts either way. That is his stated opinion, not a measured result, and it is not a licence to go collecting one-stars.
The evidence is correlational, and you should know that
None of this is proven cause and effect. The factor rankings come from an opinion survey of practitioners, and the supporting case studies are observations of single profiles with no control group. Anyone quoting you a confirmed review velocity signal with a weighting attached is making it up. Treat it as a strong hint, not a law.
What happens to your rankings when the reviews stop?
They drift, slowly, and you rarely notice until the phone goes quiet. The clearest documented example comes from Sterling Sky, where a client's reviews "flat-lined" after an internal incentive scheme ended, rankings fell, and rankings recovered once the flow restarted.
That case study, updated by Joy Hawkins in July 2025, is useful and flawed, and both halves matter. The flaw is stated in the article itself: rankings dropped around the same time as Google's Vicinity update, which changed how proximity was weighted. Two candidate causes, one outcome. Hawkins is upfront, writing that individual examples "may not serve as definitive proof of how Google's algorithm operates".
There is a second problem nobody else mentions. The mechanism was a weekly cash lottery for staff whose names appeared in reviews, which Hawkins said at the time was allowed. Google's current prohibited content policy bars merchants from asking staff to solicit reviews that identify a staff member. Read the case study for the shape of the graph. Do not copy the tactic.
How recent does a review have to be before customers stop counting it?
Sooner than most trades assume, and the window is closing. BrightLocal's Local Consumer Review Survey 2026, a panel of 1,002 US adults published in February 2026, found 74% seek reviews written in the last three months and 44% weight the last month heavily. This is US data, and no UK equivalent exists.
The number that should worry you is the two-week figure. 32% of that panel look for reviews written in the last fortnight, up from 20% the year before. Half again, in twelve months.
Now do the arithmetic on your batch. Twelve reviews collected in one week of March. By July, every one falls outside the three-month window that 74% of people care about. Your profile has not got worse. It has aged all at once, because everything on it is the same age.
One a week for twelve weeks gives you the same twelve reviews and a profile that always has something from the last fortnight on it.
Does a burst of reviews get filtered or removed?
Sometimes, but not for the reason usually given. Google's policy does not ban getting a lot of reviews quickly. It bans unusual volumes that point at manipulation, which is a different thing. Google's own explanation of why reviews go missing does not list batching anywhere.
Stated fairly, the received wisdom runs like this: collect reviews in a lump and Google's spam filter assumes they are fake and strips them out. It appears in a great many places. Almost none of them cite anything.
What Google's policy actually says
The relevant line is in Google's prohibited and restricted content policy, under rating manipulation. It prohibits:
"Content exhibiting unusual volumes or patterns of review contributions that are indicative of efforts to manipulate a place's rating"
Read the second half. The offence is unusual volume that indicates an attempt to manipulate. Volume by itself is not written as a violation. Real reviews from real customers are not manipulation, however fast they arrive.
What Google says goes wrong with missing reviews
Google's page on missing and delayed reviews lists the causes: routine policy checks that "might take a few days", recent profile merges, outdated software on the customer's phone, actual policy violations, temporary content restrictions on some profiles or categories, and reviews being disabled for schools.
Volume is not on that list. Neither is velocity. That absence is worth more than another confident paragraph. On Google's own documentation, there is no confirmed automatic filter that strips real reviews for arriving together.
Where the risk is real
The risk is not one busy week. It is a sustained abnormal pattern. Claudia Tomina of Sterling Sky published tracking data on review posting blocks in February 2026 after watching them hit legitimate businesses over roughly 60 days.
She identifies three signals that appear to combine: reviews left inside the business premises, QR codes used at the counter, and bursts of 10 or more reviews in a single day.
"Individually, none of these signals are necessarily problematic. However, when they occur repeatedly and in combination, they appear to increase the likelihood of a review posting restriction."
One tracked profile had averaged 72 reviews a month, then posted 179, 235 and 181 over three consecutive months. The block landed after roughly three months of that, and 57.5% of the reviews were deleted. In every case she tracked, the block ran the full 30 days regardless of appeal.
Keep the scale in view. A one-van plumber will never post 200 reviews in a month. The transferable lesson is about shape, not absolute numbers: a sustained departure from your own baseline, combined with on-premises collection, is what drew attention.
What a restriction looks like from your side
Google documents the consequences in its guidance on Business Profile restrictions for policy violations. A profile may be unable to receive new reviews for a set period, may have existing reviews unpublished for a set period, and may display a warning telling customers that fake reviews were removed. Google does not publish how long "a set period" is.
UK businesses carry a further layer. Under the undertaking Google gave the CMA in January 2025, UK profiles found using fake reviews face warning alerts, a deactivated review function, and for repeat offenders the deletion of all reviews for six months or longer.
The UK regulator has already named the on-off pattern
The CMA's fake reviews guidance, published in April 2025, lists things traders should not do to interfere with people's willingness to leave negative reviews. One of them, at paragraph 4.4, is interfering "by arbitrarily stopping and starting review invitations".
Be precise about what that means. The CMA is aiming at businesses who switch invitations on and off to control which reviews get published. A busy roofer who forgets to ask anyone between October and March is not the target of that paragraph.
From outside, though, the behaviour looks the same. Batch, silence, batch is hard to tell apart from a business managing its review flow for effect. A steady cadence never raises the question.
What a steady trickle looks like for a trade
Small enough that you will actually keep it up. Shaw's benchmark is the sensible one: work out how often your top competitors get new reviews, then aim for that plus one. If the three profiles above you in the map pack pick up two a month, three a month wins you the pattern.
Counting is a ten minute job. Open each competitor profile, sort reviews by newest, count how many landed in the last 90 days, divide by three. For most UK trades the answer is small. Two or three a month is a real target for a one or two van operation, and it is what your job volume supports without asking the same customer twice.
What works is not a campaign. It is a habit attached to something you already do every day: the final invoice, the handover, the moment you pack the van.
How to ask without being annoying covers the wording, and what Google and the CMA let you say.
What to do if you are starting from nothing
Be honest about the timeline. If you have four reviews, a trickle will not fix that this month, and BrightLocal's 2026 US panel found 47% will not use a business with fewer than 20 reviews.
Resist both extremes. Do not text 60 old customers on a Sunday night, and do not wait for perfect conditions either. Work backwards through recent jobs over several weeks, oldest customers last, because a review from someone you worked for two years ago reads exactly like what it is.
Frequently asked questions
How many reviews a week is safe?
Nobody outside Google knows, and no threshold has ever been published. The "10 or more in a day" figure is one practitioner's example, not a Google rule. Match your competitors' pace and you will never be near the question.
Do old reviews still count for anything?
Yes. Review count and star rating are separate ranking factors from recency, and 20 old reviews beat none. They simply stop doing the recency job, and the customer reading your profile can see the dates.
What to do this week
- Count your last 90 days. Open your profile, sort by newest, and count reviews posted since the start of the quarter. If the answer is zero, that is the whole article in one number.
- Count your competitors' last 90 days. Three profiles, ten minutes. Divide by three for a monthly pace, then add one.
- Attach the ask to a job step you already do. The invoice email, the handover, the receipt. Not a diary reminder you will ignore.
- Set a floor, not a target. One a fortnight beats twelve in March.
- Stop the annual blitz. If you have a list of old customers waiting, spread it across two months rather than two days.
- Check nothing is stacked. Several reviews going up from one postcode on one afternoon is the pattern practitioners report drawing attention.
Reviews are not a project you finish. They are more like servicing the van: unglamorous, ongoing, obvious to everyone when you have skipped it. At Business Opt the first thing we check on a trades profile is the date of the newest review, because 40 reviews and nothing since last summer tells Google and the customer the same story.
Small and constant wins.