Review monitoring
Review monitoring is tracking what customers write on rating platforms, weighing each review on its own rather than adding them up into an average. It is the only reputation channel where the customer writes the text. That weighs more than most mentions.
Published
Who it is for
First a boundary, because the word review covers more than one thing: this page is about ratings written by customers, not about reviewing software or appraising staff.
For organisations where a purchase decision hangs on a rating. That is broader than hospitality and retail: a law firm, an installer and a software vendor are all looked up first and called second these days.
And for anyone with several locations or several platforms, where it can no longer be kept up by hand and a problem at one site disappears into the average of the rest.
Why the average misleads you
A 4.2 sounds fine and says little. Someone considering you reads the worst reviews first, and those decide whether they call. One detailed two-star complaint does more than twenty five-star ratings with no text.
An average is also slow. A run of new complaints about the same thing barely moves the score, while the pattern is the signal. Steer on the average and you see that pattern only once it is too late.
And since AI assistants joined in it has a second life: assistants cite review platforms as sources. A well-written complaint sitting high on the page can end up in the answer given to someone asking about you, even if they never open the platform themselves.
What Reptor does here
We assess each review on tone and content, not only on the number of stars. A politely written three-star complaint can do more damage than an angry one-star, because it reads as more credible.
New reviews with an outspokenly negative tone trigger an alert by email, Slack or WhatsApp, so you do not first see it in the monthly report.
Because reviews sit in the same system as everything else, you also see when a review platform turns up as a source in an AI answer about your organisation. That is the moment one review starts reaching well beyond the platform it sits on, and it is exactly where review software without AI monitoring sees nothing.
Which platforms we track and what we pull
A set of profiles is ready as standard. The platforms that apply to your sector are added to those, and there is no cap on the number. Spot one we do not yet track and we will add it; that is how our coverage grows.
Per review we record the star rating, the text, the author name where the platform supplies it, the date, the language, and our own sentiment judgement with its score. Plus a link to the review on the platform itself.
How often profiles are checked is agreed with you; it depends on the plan and on how fast reviews come in. A new review with an outspokenly negative tone triggers an alert immediately either way, rather than waiting for the next round.
One thing we do that a review dashboard does not: every review is tested against what we know about your organisation. A review claiming a fact that is not true, about a price or a delivery time for instance, surfaces separately.
- "Called three times, heard nothing back"2 stars · today0.12
- "Fine product, delivery took a while"3 stars · yesterday0.44
- "Quick service, clear quote"5 stars · 2 days0.88
- Google Business Profile
- Trustpilot
- Glassdoor
Which platforms matter, and why that differs by sector
Google is the heaviest platform for almost every organisation, simply because the reviews sit next to your name the moment someone looks you up. Anyone working locally has the most to win and lose there.
Beyond that every sector has its own place. In hospitality and travel it is the booking platforms, in software the comparison sites, in professional services often a trade-specific listing or simply LinkedIn. Employer reputation is a different field again, because there it is staff writing rather than customers.
The most common mistake is tracking everything that exists. That produces a dashboard where four reviews a year on a small platform weigh as heavily as two hundred on Google. Better to establish which platforms your customers actually read, and treat the rest as background.
From individual reviews to a pattern
One bad rating is an incident, three about the same thing is a signal. Seeing that difference is the real reason to track reviews rather than read them.
So we look at more than the score. Recurring subjects surface: delivery times, a member of staff named personally, a price rise that was not explained well. That is usable information for whoever can fix the problem, and it is a different thing from a figure slipping by a tenth.
And because reviews sit beside your other channels, you see when the same subject turns up elsewhere: first in reviews, then on a forum, then in the trade press. That sequence is the most predictable route a complaint takes.
Tone of a single review
score = 0.0 to 1.0 with a class from very negative to very positive
- The score is about the text, not about the number of stars.
- score of 0.3 or below: outspokenly negative, and that is the alert threshold
- So a three-star complaint can fall below the threshold while an angry one-star sits above it.
Picture over a period
period figure = sum of scores / number of reviews in the period
- Unweighted: every review counts the same, regardless of how many people read it.
- The label beside the figure is the class that occurred most often in that period.
- This is separate from the platform's own star average, which covers the entire lifetime.
Four reviews scoring 0.8, 0.6, 0.2 and 0.9 give (0.8 + 0.6 + 0.2 + 0.9) / 4 = 0.63.
What the research says about revenue
The sharpest evidence comes from a study by Michael Luca at Harvard Business School. He compared restaurants falling just above and just below a Yelp rounding threshold: an actual score of 3.25 shows as three and a half stars, 3.24 shows as three. Those two groups are near-identical in quality and differ only in the score displayed. One star of difference proved worth 5 to 9 percent of revenue.
Two things belong with that number. It held for independent restaurants; for chains with an established name the effect was small to absent. And the study is American and about restaurants, so the figure is not a promise for a software vendor. What does transfer is the mechanism: the displayed score drives behaviour even when the underlying quality is the same.
The same effect has been found again by a different route. Anderson and Magruder looked at restaurant reservations around the same rounding threshold in The Economic Journal: half a star more made a restaurant sell out 19 percentage points more often, close to a doubling against the baseline. Two independent studies, the same direction, the same trick for holding quality constant.
More useful for daily work is what a meta-analysis in Journal of Retailing showed. Floyd and colleagues pooled twenty-six studies covering more than four hundred measured relationships and found that the tone of reviews relates more strongly to sales than the number of reviews does. What people write weighs more than how many people write. Chevalier and Mayzlin had found the same in book sales: a one-star review did more than a five-star one.
On how many people act on it, BrightLocal's Local Consumer Review Survey is the most cited source. In the 2026 edition, among 1,000 US consumers, 97 percent said they read reviews before choosing a local business, 41 percent said they always do, and 77 percent said negative reviews put them off. Nearly a third looks no further below 4.5 stars.
Review monitoring or review management?
These two words get used interchangeably and they do not mean the same thing. Review management is the whole business of handling ratings: collecting new reviews, replying to them, syndicating them to your site, and sometimes requesting a removal. Review monitoring is the part that establishes what is there and what it means.
Anyone searching for review management software is usually after the first: a system that asks for ratings and streamlines replying. That is a different product with a different price and a different setup, and there are firms that do it well.
We do the second, deliberately and completely. It is not a reduced version of the first; it is the part that is about your reputation rather than your workflow. Where review management wants to improve your ratings, monitoring shows what they are doing right now, and where the same subject surfaces on other channels.
In practice that means the reply to a complaint is written by you or your agency. Answering takes tone, knowledge of the case and sometimes a lawyer, and a recognisably automated reply to a fair complaint is worse than no reply. Collecting new ratings is not something we do either; that is its own discipline, and most platforms place conditions on how you may ask.
| What it is about | What it does | Do we do this? | |
|---|---|---|---|
| Review monitoring | Your reputation | Establishing what is there, what the sentiment is, and where the same subject surfaces elsewhere | Yes |
| Responding | Your workflow | Writing replies to complaints and compliments | No, you or your agency write those |
| Collecting | Your workflow | Asking customers for a rating, within each platform's rules | No, that is a craft of its own |
About the figures
How we determine the tone of a single review, and how those scores add up to a picture over a period, is set out in the measurement protocol.
How we calculateSources
The research this page refers to, with its scope attached. A figure from a US consumer survey is not the same as a figure about a business market, and that belongs on the page.
- 01
- 02
- 03
- 04
- 05
What Reptor offers here
Questions about the offering on this subject: what you get, what it costs, and where we stop.
Which platforms do you track?
The platforms that apply to your sector, from the general rating sites to wherever your field actually writes. Which those are is established during setup, because it varies: a hotel has a different list from a software vendor.
Do I get an alert for a bad review?
Yes, as soon as a review arrives with an outspokenly negative tone, by email, Slack or WhatsApp. You set the threshold yourself.
Can you help remove an unfair review?
That is not part of the monitoring. We establish what is there and how it sits against the rest; a removal request or a legal route is different work.
Does this work across multiple locations?
Yes. Each location can be tracked as its own entity, with its own picture beside the total. That is also the reason to do it: in a single average across all locations, a problem at one site disappears.
What happens to older reviews?
They are included during setup, so you have a starting point rather than waiting months for a line to appear. History stays available for a year.