How often a new review has to arrive
Stop counting your total and start counting your month. 74 percent of consumers want reviews from the last three months and 44 percent look for the last month, so the number that matters is how many arrive per month, not how many you have. Three a month keeps nine in the window that gets read.
Why a total is the wrong number
A business with 200 reviews and nothing since spring reads worse than one with 30 reviews and four from this month. The reason is in the data: 74 percent of consumers seek reviews written in the last three months, and 44 percent pay attention to the last month.
That turns the whole subject from a stock into a flow. What is visible in the window that people read is your rate multiplied by three, and nothing else. The back catalogue still counts for the average, it just does not count as evidence that you are still trading well.
Working out your own number
The arithmetic is one line. Reviews visible in the three month window equal your monthly rate times 3. At 1 a month the window holds 3, at 3 a month it holds 9, at 5 a month it holds 15.
Now spread it. Consumers consult an average of six review sites, and the platform mix has shifted: Google's share as a source fell from 83 percent to 71 percent while Apple Maps went from 14 percent to 27 percent in one year. Three reviews a month landing entirely on one platform leave the other five sites looking abandoned.
Then add the response side. 19 percent of consumers expect a same day response, so a steady three a month is also a commitment to roughly 3 replies a month written within a day. That is fifteen minutes of work, and it is the part most often dropped.
| Reviews per month | Visible in the window | Asks needed at one in four |
|---|---|---|
| 1 | 3 | 4 |
| 2 | 6 | 8 |
| 3 | 9 | 12 |
| 5 | 15 | 20 |
| 8 | 24 | 32 |
How to hold a rate without begging
Tie the ask to a moment that already exists: the handover, the final invoice, the follow up call. A fixed moment produces a stable rate, while a campaign produces a spike that decays and leaves the window empty three months later.
Ask a fixed share rather than everyone. If one in four asked customers writes something, then twelve asks a month produce your three. Write that number down, because it is the only part of the process you control.
Questions and answers
Do old reviews stop counting?
Is a burst better than a trickle?
How many platforms should the rate cover?
What about replies?
Sources
- BrightLocal, Local Consumer Review Survey 2026 Recency expectations, response expectations and platform shares
- Google Search Central, Local business structured data How review and rating data is expressed for machines
