What Is Review Velocity?
Review velocity is the rate at which a business earns new reviews over time, and the shape that rate makes across weeks and months. It is not your star rating.
Review velocity is the rate at which a business earns new reviews over time, and the shape that rate makes across weeks and months. It is not your star rating and it is not your review total. Two businesses can hold the same average from the same number of reviews and still look nothing alike once you plot when those reviews arrived.
Why the rate matters more than the total
A review total is a historical fact. A review rate is evidence about the present: whether the business is still trading, still serving customers, and still doing it well enough that people bother to say so. That is the question a local search result has to answer, and a four-year-old cluster of reviews answers it badly.
The practical failure is a stall. You collect reviews hard for a quarter, hit a number that feels respectable, and stop. Nothing appears to break. Then a competitor with fewer reviews and a slightly worse average starts outranking you locally, because their profile shows a business that is visibly alive and yours shows one that peaked eighteen months ago. Rebuilding a review habit takes longer than maintaining one, which is a large part of why steady local work is priced monthly rather than as a one-off project.
How platforms read the pattern
Reviews are not simply counted on arrival. Each one carries metadata, and the metadata is assessed before the review is shown, kept or quietly withheld.
- Timing is stored. Every review has a timestamp. Once you have a timeline, gaps and clusters become visible without anyone needing to look for them.
- Your own history sets the baseline. The comparison that matters most is you against you. A rate that looks unremarkable for a busy restaurant looks strange for a solicitor who has averaged one review a quarter for three years.
- Provenance is checked alongside timing. Account age and history, review text, and whether a batch of reviewers has anything else in common. A burst is far more suspicious when the accounts behind it are new and otherwise inactive.
- Filtering happens silently. Google removes policy-violating and fake reviews, and it does not notify you when a review is withheld. Reviews can vanish days after they appear, and you will not receive an explanation for each one.
- The record attaches to the entity. Reviews sit against the business record, not your website, which is why review work and site work are managed separately even though they compete for the same rankings.
All of this runs against the profile customers actually see, so an unclaimed or badly maintained listing undermines review work before it starts. Verification, categories and opening hours are not separate from the review question. They are the context the pattern is read against.
| What the timeline shows | How it tends to read |
|---|---|
| A steady trickle roughly matching your trading volume | Ordinary. This is the pattern you want and the least interesting to a filter |
| Months of silence, then a tight cluster over a few days | The shape most associated with solicited batches and bought reviews |
| Reviews arriving faster than you could plausibly serve customers | A mismatch between claimed activity and observable activity |
| A cluster written by accounts with no other review history | A provenance problem, independent of the timing |
| A modest lift after a genuine campaign, then a return to baseline | Normal, provided the underlying customer volume supports it |
The blunt version
Silence followed by a burst is the exact shape review filtering exists to catch. It is also the shape almost every review campaign produces, because campaigns start when someone finally notices the profile looks thin. You then send one email to a year of past customers, twenty reviews land in a week against a baseline of roughly none, and you have generated the most flagged pattern in local search using entirely real customers.
Now the part the tool vendors leave out: Google publishes no threshold. There is no documented number of reviews per day, per week or per month that triggers a filter, and any agency quoting you one has invented it. We would rather admit the number does not exist than sell you a rule we made up. What is documented is that fake and policy-violating reviews are removed, and that removal happens without notice.
The workable rule needs no number. Ask for reviews continuously, from customers as they are served, at whatever rate your actual transaction volume produces. Steady beats spiky, not because steady is rewarded, but because spiky is the thing that gets examined. It is also slower to show results, which is worth knowing when you read what the first months look like on any local engagement.
Example
Say a dental practice has collected reviews sporadically for two years, averaging a handful a quarter. A new practice manager exports the patient list and emails everyone at once. Thirty-one reviews arrive in nine days, all genuine, all from real patients. Within a fortnight several have disappeared from the profile with no notification and no way to appeal individually. The practice concludes that reviews do not work. What actually happened is that a two-year baseline of near-silence was interrupted by a spike no ordinary trading pattern would produce. The same thirty-one requests, sent to patients at the end of each appointment across six months, would have looked entirely ordinary.
FAQ
How many reviews a week is safe to collect?
Google publishes no threshold, so nobody can honestly give you a number. The useful test is plausibility against your own trading: a rate your customer volume could genuinely produce, sustained rather than concentrated. A busy salon and a commercial surveyor have very different plausible rates, and both are fine.
Why did reviews disappear from my profile?
Google removes reviews that breach its policies, including fake and incentivised ones, and it removes them without telling you which or why. Genuine reviews are sometimes caught alongside them, particularly when they arrive in an unusual cluster. There is no per-review appeal that reliably reverses this.
Can I offer a discount for leaving a review?
No. Google prohibits incentivised reviews, and offering anything of value in exchange breaches that policy whether or not the review is honest. Asking is allowed and encouraged. Paying, discounting or running a prize draw for reviews puts the whole profile at risk, not just the incentivised reviews.
Related terms
- Google Business Profile — the listing your reviews attach to, and the thing that has to be right first.
- ccTLD — location signalling at country level, for businesses whose problem is borders rather than streets.
- Product Feed — the equivalent freshness problem for retailers, where stale data quietly removes you from view.
Nobody outside Google knows the threshold, because Google has never published one. What is knowable is the shape: silence, then a spike, is the pattern worth avoiding. Collect reviews the way you serve customers, one at a time, forever.