The core metrics
| Metric | What it tells you | How to work it out |
|---|---|---|
| New reviews per month | Whether your request process is running | Count of reviews received in the month, by platform |
| Recency | Whether your page looks active to a new customer | Days since the most recent review |
| Rating distribution | Whether problems are rare or common | How many 1, 2, 3, 4 and 5 star reviews, not just the average |
| Response rate | Whether you're keeping up | Reviews with a reply divided by all reviews in the period |
| Response time | How quickly unhappy customers hear back | Typical days between a review and your reply |
| Request-to-review rate | How well your request message works | New reviews divided by requests sent, from the same period |
| Recurring themes | What to fix in the business | Tally the topics mentioned: waiting, price, cleanliness, staff, quality |
Why the average rating misleads
The average moves slowly once you have many reviews, so it hides recent change. It also hides shape: a page with mostly fives and a cluster of ones tells a different story from one full of threes. Look at the distribution and at the last 90 days separately from your all-time figure.
Don't chase a perfect score either. A page with only five-star reviews can look less believable to customers than one with a few honest criticisms and good replies.
Themes are the most useful number
Once a month, read the reviews from that month and tally the topics. If "waiting" comes up in a quarter of complaints, that's an operations problem no reply template will fix. Software with topic analysis can speed this up, but a spreadsheet works for a small business.
Numbers to treat with care
- Profile views and impressions. Useful context, but they depend on search behaviour as much as your reviews.
- Composite "reputation scores". Vendors calculate these in different ways. Fine for tracking your own trend in one tool; meaningless for comparing across tools.
- Competitor ratings. Interesting, but you can't verify how they get their reviews. Compare your own trend over time instead.
- Sentiment percentages. Automated sentiment can misread sarcasm and mixed reviews. Spot-check before trusting it.
A monthly review routine
- Pull the numbers. New reviews, distribution, response rate and response time for the month, by location.
- Read every negative review. Tally themes and note anything that needs a fix.
- Check your request flow. Did requests go out to every customer? Did any integration break?
- Agree one change. One operational fix and one process fix, owned by a named person.
- Write it down. A short monthly note makes the trend visible over a year.
Never set staff targets for star ratings or review counts that encourage pressure on customers, filtering who gets asked, or reviews written by staff. Target the process (requests sent, replies on time) instead.
To set up this routine from scratch, see the 30-day plan.