Reviews are the part of local SEO clients ask me about most, and the part where the advice online runs from slightly out of date to actively illegal. Google is unusually open about the fundamentals here. Its own local ranking documentation says review count and review score factor into local ranking, so the question was never whether reviews matter. The question is which parts of the reviews work move anything, and which parts are wasted effort or, since last year, a legal risk.
This piece covers the three conversations I have most often. Why a steady trickle of reviews beats a big burst, what replying actually does and doesn't do, and where the line now sits, because since April 2025 the UK has real law on reviews, and since March 2026 the regulator has been using it on household names. Google tightened its own policy in April 2026 as well, so the rules are stricter on both fronts than when most review advice was written.
Where reviews sit in the ranking contest
The map pack runs on relevance, distance and prominence, and reviews are the loudest signal in the prominence bucket. Count and score are the confirmed inputs. A business with 150 reviews at 4.7 is, all else being equal, more prominent than one with 12 at 4.9, and both beat a profile sitting on four reviews from 2022.
All else is rarely equal, though, and reviews won't rescue a profile with the wrong primary category or a website that never mentions the town you serve. I've covered that side of the contest in my guide to Google Business Profile optimisation, and reviews work best as one strand of the wider local SEO effort rather than the whole plan.
There's a useful side effect worth knowing about too. Customers describe what you did in their own words, the boiler replaced on a Sunday, the wedding cake delivered in a heatwave, and that text gives Google more to match against real searches. I treat that as a bonus rather than something to engineer. Review text also feeds the AI assistants that now answer "who's the best plumber near me" searches, which read reviews as third-party corroboration (I cover that in my guide to generative engine optimisation). None of that justifies scripting keywords into customers' mouths, and since April 2026 Google's rating manipulation policy explicitly bans asking for reviews that include particular content or an employee's name.
Velocity, or why the trickle beats the burst
The most common pattern I see is the January review push. Someone reads that reviews matter, emails every customer from the last five years, gains thirty reviews in a fortnight, and then the profile falls silent again until somebody remembers eighteen months later.
That pattern underperforms for two reasons. The first is recency. The local ranking factor surveys keep reporting what practitioners see on the ground, which is that recent reviews carry more weight than old ones, and a pile gathered in one month ages together. By summer, your January push is stale and the competitor collecting four a month has quietly gone past you.
The second is the filter. Google holds back reviews its systems find suspicious, and a sudden spike on a previously quiet profile is exactly what a bought batch looks like from the outside. I've seen legitimate reviews from real customers vanish into the filter during over-enthusiastic pushes. They don't come back, and there's no appeals desk worth the name.
In short, match the ask to the rhythm of the business. Build it into the end of every job, every delivery, every completed project, so reviews arrive at the pace custom does. A plumber doing five jobs a week should see a few reviews a month, every month. That cadence looks normal to Google's spam filters, and every month adds another recent review to the pile.
Replies, and what they're actually for
Google explicitly encourages replying to reviews, and its guidance links responding with visibility. Whether replies are a direct ranking input has never been confirmed, and I'm not going to pretend otherwise. I reply to every review on the profiles I manage anyway, for three reasons that don't depend on the algorithm.
Future customers read them. The reply to a negative review is the most-read text on your profile, because it's where people find out what you're like when something goes wrong. A calm, factual reply that acknowledges the problem and states what you did about it wins business from the exact people the review was supposed to scare off.
Reviewers notice them. People are measurably more willing to leave a review for a business that visibly reads and answers them, so replies feed the velocity point above.
And they're your only voice on that part of the page. Everything else in the reviews panel is written by other people. The reply is the one bit you control, which is a strange thing to leave blank.
Two rules for the negative ones. Reply quickly, within a day or two, and never argue. The audience is the hundreds of people who'll read the exchange, not the one person who wrote it. Take the detail offline, fix what's fixable, and let the reply show the pattern.
Asking properly
Everything above depends on actually asking, and most businesses simply don't. The ones that do well share the same habits.
They ask everyone, not a filtered shortlist of the delighted, and they don't set staff a review quota or tell them what to say. Google added both to its rating manipulation policy in April 2026, which matters most for multi-site operators, where one badly designed incentive scheme becomes a violation at every branch (more in my multi-location SEO guide). They ask at the moment the value lands, when the job's signed off or the results are in, not three weeks later by newsletter. They make it effortless, with the profile's direct review link in the follow-up email or a QR code on the invoice. And the ask comes from a person, because "it would really help us out" from the tradesperson at the door converts far better than an automated platform email ever will.
What you must not do is screen first. The "how did we do?" email that routes happy customers to Google and unhappy ones to a private feedback form is called review gating, and it's explicitly against Google's rules, which prohibit selectively soliciting positive reviews. The regulator takes the same view of practices that suppress negative feedback. Ask everyone, take the occasional two-star on the chin, and let the profile look like what it is, a real business with a real spread of customers.
The line, in a table
Since 6 April 2025, the Digital Markets, Competition and Consumers Act has turned most of the old grey areas black and white in the UK. The CMA published guidance the same month, and on 27 March 2026 it opened its first investigations under the new rules into five businesses. Autotrader and Feefo are being looked at over whether one-star reviews were excluded from ratings, Dignity over whether staff were told to write positive reviews, Just Eat over how its rating system treated star ratings, and Pasta Evangelists over undisclosed discounts for five-star reviews. The CMA said it expected to give an update in September 2026, and I haven't found a published outcome yet, so treat these as open investigations rather than findings. The maximum penalty is £300,000 or 10% of global turnover, whichever is higher, and individuals can personally be fined up to £150,000.
Here's where the common tactics now stand.
| Tactic | Google's rules | UK law since April 2025 |
|---|---|---|
| Asking every customer for an honest review | Encouraged | Legal |
| A direct review link or QR code | Encouraged | Legal |
| Gating: screening out unhappy customers first | Banned | High risk, suppression practices are in the CMA's first investigations |
| Discounts, freebies or prizes for reviews | Banned | Banned unless the incentive is clearly disclosed, and the CMA is investigating exactly this |
| Reviews from staff, family or your agency | Banned | Banned, these are fake reviews (staff-written reviews are one of the CMA's investigation topics) |
| Review quotas for staff, or asking for reviews that name an employee | Banned since April 2026 | Not named in the CMA's announcement, but it sits close to the staff-written reviews under investigation |
| Buying reviews | Banned | Banned, with the largest penalties |
The pattern is that the law now roughly matches what Google's policies said all along. The difference is enforcement. Google's penalty was removed reviews, a suspended profile or, in recent cases, a public warning banner on the listing. The CMA's penalty is a percentage of turnover with your company's name in the press release.
Fake negatives, briefly
The mirror problem is the review you didn't earn, from a competitor, a never-customer, or someone confusing you with another business. Flag it through the profile, be specific about which policy it breaches, and be patient, because removal takes days to weeks when it works at all. The new law helps here too, since platforms now have a duty to take reasonable and proportionate steps against fake reviews, and removals have visibly improved since it landed. While you wait, reply calmly stating you have no record of the customer. Readers are better at spotting a fake than you'd expect, and far better than the filter.
If you run location pages as well, the reviews that name an area and a job are the ones worth building those pages around, and I cover how in local landing pages that rank. The fuller playbook, citations, links and the rest, is in my local SEO best practices guide.
Google reviews and SEO FAQs
Do Google reviews actually improve rankings?
Yes, and this one isn't speculation, Google's documentation states that review count and score factor into local ranking as part of prominence. What reviews can't do is compensate for the wrong categories, a thin profile, or a website with no local relevance. They're a multiplier on a sound setup, not a substitute for one.
How many Google reviews do I need?
There's no threshold, because the contest is relative. Look at the three businesses currently holding the map pack for your main search and compare their counts, scores and recency with yours. That gap is the target. For most small service businesses, a steady few reviews a month closes it within a year, and I'd take that pattern over a one-off blitz every time.
Can I offer customers a discount for leaving a review?
No. Google bans incentivised reviews outright, and since April 2025 UK law bans them wherever the incentive isn't clearly disclosed. The CMA's first investigations under the new act include businesses offering discounts for five-star ratings, so this has moved from frowned upon to genuinely dangerous. If you want more reviews, improve the ask, not the reward.
Should I reply to negative reviews?
Yes, quickly and calmly, and treat the reply as written for future customers rather than the reviewer. Acknowledge the issue, state the facts without arguing, and move the detail offline. A well-handled negative review sitting in public does more for trust than a suspiciously perfect wall of five stars.
Want the local SEO handled properly?
Tell me what you're working on and what you're trying to achieve, and I'll give you an honest view of whether I can help and what it would take.
Get in touch