How review velocity affects your local rankings

What review velocity actually is

Velocity is two numbers, not one. The first is your rate: new reviews per month. The second is your consistency: how much that rate varies month to month. Businesses obsess over the first and ignore the second, which is backwards, because the second is where the damage usually happens.

Two businesses can finish a six month stretch with an identical review count and be in completely different positions.

What is published, and what is inference

This distinction matters more than anything else in this guide. Local ranking is not an open system. Be suspicious of anyone, us included, who states a weighting as fact.

Published by the platforms

Informed inference, stated as inference

Why a steady trickle beats a burst

There are four separate reasons, and they stack.

1. Filtering risk is concentrated in bursts. Spam systems look for anomalies. A rate that has been four a month for a year and then hits forty in a week is an anomaly by definition. Steady growth never presents that shape.
2. Humans only read the top of the list. A prospective customer scans maybe five to ten reviews. If your most recent is from last week, your business reads as alive. If it is from last November, the reader wonders what happened.
3. Continuous inflow dilutes bad reviews continuously. A one-star landing on a profile that adds five reviews a month is one voice among many within a fortnight, and it moves the average barely at all. On a dormant profile it is a large share of everything a reader will see, and it stays that way until you do something about it.
4. A cadence is a process. A burst is an event. Events end. Every business that has run a review push knows the feeling of the number going flat again in month two. The process version survives staff turnover, because it is attached to job completion rather than to someone's enthusiasm.

How recency weighting actually works

Recency operates on two layers, and only one of them is about algorithms.

Layer one: what the reader sees

Be careful with the common version of this claim. Google's default sort on a Business Profile is Most relevant, not Most recent, so freshness does not automatically buy you the top slot, and readers who do switch the sort are a minority. What is reliable is the second half: a reader scans a short run of reviews and stops. Your effective reputation is not your lifetime average. It is roughly the handful of reviews a stranger actually reads, plus the star number next to your name. That is why a business with a 4.6 across 300 reviews can still lose the call to a 4.4 across 60, if a reader who does check dates finds nothing newer than eighteen months.

Layer two: what the ranking systems do with it

Google publishes that review count and score feed prominence. It does not publish a decay curve. The reasonable working assumption, held loosely, is that a profile receiving reviews continuously is a stronger signal of an operating, in-demand business than an identical count that stopped accumulating two years ago. Treat that as a sensible bet, not a law.

A practical heuristic that costs you nothing if the inference is wrong: treat reviews older than twelve months as decoration. They pad your count and reassure nobody. Plan your cadence so that at any point in the year you have a respectable number of reviews from the last ninety days.

Why a spike looks manipulative, platform by platform

The reason to avoid spikes is not superstition. Each platform has a specific mechanism, and they differ enough that a single strategy across all of them is a mistake.

Google

Google removes policy-violating reviews, sometimes days or weeks after they post, and it removes them in sweeps. The classic own-goal is the front-desk QR code: fifteen customers in one afternoon, on the shop wifi, several creating a Google account for the first time in order to post. Every one of those signals is individually innocent and collectively looks synthetic. The fix is trivially easy. Send the link instead, and let people review from their own phone on their own network, hours later. Our Google review link and QR generator builds the direct link for exactly that purpose. Print the QR if you like, just do not stand over people while they use it.

Yelp

Yelp is the exception to almost every piece of review advice. It prohibits soliciting reviews outright, and its recommendation software routinely moves reviews into the not-recommended section, where they do not count toward your rating. First-time reviewers arriving in a cluster is close to the archetype of what that software is looking for. On Yelp your velocity plan is: do not have one. Claim the page, keep the information accurate, respond to what arrives, and put the badge on your site.

Facebook

Recommendations are attached to real, visible profiles, which shifts the risk from algorithmic filtering to human judgement. A reader can click through and see that six recommendations came from accounts with no photos and no history. Facebook may leave them up. Your prospective customer will still discount them.

The vertical platforms are their own case, and the table at the end of this guide covers them alongside the three above rather than repeating them here. The short version: they moderate more slowly and more manually, and a rejected batch is hardest to appeal there.

Where response rate fits in

Response rate is the percentage of your reviews that have an owner reply. It is the one number on this page you can take to 100 percent this month, for free, using nothing but time.

Google encourages owners to respond and says it builds trust with customers. It has not published a ranking weight, so treat the direct-ranking benefit as unproven. The indirect benefits are not speculative:

• A reader deciding between you and a competitor sees an owner who answers, including on the bad ones. That is a conversion effect on the same page that ranking would have delivered.
• Answered negatives read as resolved incidents. Unanswered negatives read as admissions.
• Responding raises velocity. Customers who see the owner replying are more likely to think a review will be read, so more of them write one.
• Some platforms describe responsiveness as part of how they score a listing. ApartmentRatings presents manager responsiveness as an ingredient in its epIQ grade, for example. How much it counts is not clearly documented publicly, so treat that as a stated ingredient rather than a known weighting you can optimise against, and check the platform's own current documentation before you plan around it.

Practical rule: respond to everything within 48 hours, positives included, and keep replies short and specific. If drafting is the bottleneck, the free review response generator will get you a first draft that sounds like an owner rather than a legal department. Edit it before you post it.

A cadence a real business can sustain

This is the part that survives contact with a busy week. Work through it in order.

1. Measure your baseline. Count reviews received in the last 90 days and divide by three. That is your current monthly rate. Write it down. Everything else is relative to this number.
2. Set an increment, not a multiple. If you are at three a month, target five or six. Do not target thirty. A modest increase held for a year beats an aggressive target abandoned in March, and it never produces the burst shape.
3. Convert the target into asks. You need a placeholder conversion rate to size the effort, so use one in ten: ten asks, one review. Six reviews a month therefore means about sixty asks. Be clear about what that number is. It is an illustrative planning figure, not a benchmark and not a promise. Real ask-to-review conversion varies enormously by trade, channel, timing and how well the customer knows you, and plenty of businesses land well below one in ten. Anyone selling you a headline figure like one in five is quoting their best case, not a conservative one. Measure your own after 60 days and throw the placeholder away.
4. Attach the ask to an event, never to a calendar reminder. Job closed, visit completed, ticket resolved, meal finished. This is what makes the cadence self-spacing: your work is already distributed across the week, so your asks will be too.
5. Cap the daily volume. Set a ceiling of roughly twice your normal daily rate. If a big week produces more completed jobs than that, queue the overflow into the following days rather than sending them all at once.
6. Split across platforms deliberately. Send most customers to Google, some to the platform that matters in your vertical, and none to Yelp. Rotating also keeps any single profile from showing a suspicious spike.
7. Review it monthly, on a fixed date. Four numbers: reviews received, asks sent, response rate, and average rating of the last 30 days. Fifteen minutes. If reviews received dropped, the cause is almost always that asks sent dropped first.

The per-platform rules in one place

Platform policies change. Check the current review policy for each platform before you build a process on it. Primary sources for the rows above: Google prohibited and restricted content, Google local ranking guidance, and the Yelp Content Guidelines. For Facebook and the vertical platforms, consult each platform's own current published policy directly. The regulatory text referenced on this page is the Federal Trade Commission's Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465.

Recovering after a bad stretch

A rough month, a staffing problem, one review that went around a local Facebook group. The instinct is to flood the profile with positives immediately. That instinct produces exactly the burst shape that gets filtered, at the exact moment your profile is under the most human scrutiny.

Do this instead, in order.

1. Respond to every negative review first, within a day, and make the responses specific about what changed operationally. Readers weigh a substantive reply heavily.
2. Fix the underlying cause before you increase asks. Higher velocity on a broken process just produces more one-stars, faster.
3. Raise the cadence by a modest step, roughly 50 percent above your baseline, and hold it for a full 90 days. Slow recovery is the only recovery that sticks.
4. Dispute only reviews that genuinely violate platform policy: not a customer, conflict of interest, off-topic, or clear policy breach. Disputing a fair negative wastes credibility you will want later.

Five mistakes that quietly kill velocity

Gating deserves the extra warning, because most of what you will read about it is wrong in one direction or the other. Two separate questions:

• Platform policy is settled. Google's review policies prohibit review gating, meaning selectively soliciting reviews only from customers you expect to be happy. That is not an interpretation, it is what the policy says.
• The regulatory position is narrower and less settled. The Federal Trade Commission's 2024 Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, does not name gating as a per se violation. The Commission declined to codify an express prohibition. That is not affirmative permission: review suppression can still be reached under 465.6 and under Section 5, depending on the facts. Nobody should state this as settled certainty, in either direction.

Which makes the defensible practice easy to describe. Ask every customer. Offer a private feedback channel in addition to the public ask, never instead of it. We set out where the lines sit on our FTC compliance page.

Where software fits, and where it does not

The first pass takes an afternoon, and it is steps one to three of the cadence above: baseline, increment, ask trigger. Everything after that is the same three things repeating, which is the point.

You do not need software for any of this. A spreadsheet, a saved link, and a calendar reminder will run the whole cadence, and for a single-location business that is often the right answer. Software earns its place when the asks have to fire automatically after each job, across several platforms and locations, without anyone remembering. If that is where you are, Praisly starts at $59 a month, month to month, with every price published and no sales call. Home services in particular lives and dies on this loop, and the trade-specific version is on the HVAC, plumbing and electrical page.