Reviews · · 7 min read
Google reviews analysis: from star ratings to a fix list
How to analyse Google reviews properly: read the text, not the stars, group mentions into topics, track tone and share over time, and turn it all into a fix list.
For most local businesses, Google is where reviews matter most. It sits next to your opening hours and your directions, so a customer deciding between you and the place down the road reads your reviews at exactly the moment they choose. Yet most teams treat Google reviews as a score to protect rather than a source of information.
This guide covers how to analyse Google reviews so that they produce a short, ranked list of things to fix, and how to keep that list honest over time.
What makes Google reviews different
Google reviews have a few features that change how you should read them.
- They are short. Many Google reviews are a sentence or two, and plenty are a star rating with no text at all. You need more of them to see a pattern than you would on a site where people write paragraphs.
- Recency matters to readers. People scanning your profile mostly read the newest reviews. A problem from three years ago matters less to them than a run of complaints from the last month.
- They are public and local. Your profile appears in local search and on maps, so your reviews are part of the first impression for people who have never heard of you.
- Your replies are visible. Owner responses sit under each review. Future customers read them as evidence of how you treat people when something goes wrong.
- They span every location. If you run several sites, each has its own profile, its own rating and its own set of reviewers. Comparing them is where much of the insight hides.
None of this changes the basic method. It changes what you prioritise and how often you look.
Why the star rating is not the analysis
A 4.4 average tells you that most people are reasonably happy. It does not tell you why, or what to do next. Three common situations show the problem:
- The rating is stable but the reasons are shifting. Customers start praising your new booking system and complaining about parking. The two cancel out and the average does not move.
- A small problem is growing. A complaint that appears once a month barely touches the average. By the time it shows up in the rating, it has been bothering customers for a season.
- Five stars, mixed text. "Great food, shame about the wait" often comes with four or five stars. The rating records the praise and drops the complaint.
The useful information is in the text. The job is to read it in a way you can count.
Step 1: Gather the full history
Export or import every review you can, not just the latest page. For each one, keep:
- the date
- the star rating
- the full text, including any text added when the reviewer edited it
- the location, if you have more than one
- your reply, if you wrote one, and when
Reviews with stars but no text still count towards your rating, so keep them for volume and averages. They simply contribute nothing to the topic analysis.
If you also run post-visit surveys or a feedback widget, put those replies next to your Google reviews. Public reviewers are a self-selected group; your own survey replies tell you what the quieter majority thinks. The posts on TripAdvisor review analysis and combining Trustpilot and Yelp reviews explain how to bring other sources into the same picture.
Step 2: Tag every mention with a topic
A topic is one thing customers talk about. For a café it might be coffee, food, service speed, staff, prices, seating and cleanliness. For a dental practice it might be booking, waiting time, reception, the dentist, pain and comfort and cost.
Tag every topic in a review, not only the main point. "Lovely staff, but the wait was forty minutes and the coffee was cold" mentions three topics.
Let topics nest. Service is a starting point, but service → waiting time and service → order accuracy are what someone can act on. Start broad and split a topic when it grows too large to act on as one thing. Merge topics that turn out to mean the same thing.
Step 3: Give each mention its own tone
Mark each mention as positive, negative or mixed. In the example above, staff is positive, waiting time and coffee are negative, all in one review that might carry four stars.
This is the step that turns a pile of reviews into something you can sort. Once every mention has a tone, you can see which topics customers love, which cost you stars, and which split opinion. For more on tone and how to keep it reliable, see customer feedback sentiment analysis.
Step 4: Measure change as share, not volume
Review volume on Google rises and falls with your trade. A busy December brings more reviews, so every topic looks like it grew. Counting mentions will mislead you.
Use share of conversation instead: the percentage of reviews in a period that mention a topic. Then look at its tone.
| Topic | Last month | This month | Tone this month |
|---|---|---|---|
| Waiting time | 9% of reviews | 17% of reviews | Mostly negative |
| Staff | 31% of reviews | 30% of reviews | Mostly positive |
| Parking | 3% of reviews | 8% of reviews | Mostly negative |
These figures are invented for illustration. Read this way, a hypothetical restaurant would see that waiting time has nearly doubled its share and turned sour, that staff remain a steady strength, and that parking is new and worth a look.
Step 5: Turn topics into a fix list
Rank the negative topics by three questions:
- How many customers does it touch? Its share of conversation.
- How badly does it hurt? How often it appears in one- and two-star reviews.
- Is it getting worse? Its change against the previous period.
Common, painful and growing goes to the top. Rare and stable can wait, even if the reviews that mention it are the angriest on the page.
For each of the top three topics, read ten real reviews before deciding anything. The numbers tell you where to look. The text tells you what is actually going on: whether "slow service" means the kitchen, the till or the table turnover on Saturday lunchtime.
Write the fix list in plain terms, with an owner and a date:
- Add a second person on the till from 12 to 2pm on weekends. Owner: floor manager. Check in four weeks.
- Put parking directions on the booking confirmation. Owner: whoever manages bookings. Check in four weeks.
Step 6: Compare locations
If you run several locations, analyse them side by side using the same topics. Raw averages across sites are hard to compare because each has a different mix of customers and a different number of reviews. Share of conversation and tone per topic are much easier: if cleanliness is negative in 4% of reviews at three sites and 15% at the fourth, you know where to send someone first.
Comparison also shows you what a good site does differently. If one location is consistently praised for staff, find out what its manager does and copy it. The guide on comparing feedback across locations goes into this in more depth.
Step 7: Use your replies as part of the fix
Owner replies are not just damage control. A short reply that names a specific change shows future readers that someone listens, and it gives the original reviewer a reason to come back. Keep a note of which topics you replied to and what you promised, then check whether those complaints actually fall afterwards. Our guide on responding to negative reviews covers how to write replies that help.
Step 8: Keep the cadence
A one-off analysis goes stale quickly. A light monthly routine works for most businesses:
- Bring in the month's reviews.
- Check which topics grew or shrank in share and which changed tone.
- Re-read the reviews behind anything that moved.
- Update the fix list and check last month's fixes.
If you also ask customers for reviews, make sure you ask everyone, not only the happy ones. That keeps your Google reviews a fair sample and keeps you within Google's rules. See how to ask customers for reviews.
Doing this with Userforms
Userforms imports Google reviews alongside TripAdvisor, Yelp and Trustpilot on the Plus plan and above, and the Pro plan can sync directly with your Google Business Profile. Each review is read as it arrives: filed into a topic hierarchy that you can rename, merge or split, with a tone for each mention.
The dashboard's What changed briefing shows which topics grew or shrank in share of conversation and which turned better or worse against the previous period, and every chart opens onto the reviews behind it. If you run several sites, you can compare locations side by side, and each of your feedback forms writes its own report: the short answer first, then the findings with the chart and replies behind them.
You can start for free, or see what each plan includes on the pricing page.