What an average rating hides when a market has only a few places
In a town with eight dentists, one bad score can drag the average down. Why small markets need counts, spread and review numbers beside the average, and how to read them.
Part of the guide Location intelligence automation
In a small market, an average rating is mostly a story about one or two places. With eight businesses, a single poorly rated one can pull the average down by a third of a star, and a newcomer with three glowing reviews can push it up just as far. Read the average only beside three other numbers: how many places have a rating at all, how many reviews each one has, and how far apart the best and worst are.
This article is for owners weighing a move into a small town or suburb, franchise teams comparing territories, and analysts who report average ratings to a committee. It shows how few places it takes to move an average, a worked example, the misreadings that follow, and how an AI employee can give you the fuller picture automatically.
Why small numbers swing so much
An average treats every place as one vote. In a city with hundreds of cafes, one bad score barely moves it. In a market of eight, each place is an eighth of the answer. Plain arithmetic shows the effect: seven places rated 4.6 and one rated 2.0 give an average of about 4.3. Remove the one outlier and the average is 4.6. Nothing about the other seven changed.
Review counts add a second layer. A 5.0 from four reviews and a 4.4 from four hundred are not the same kind of evidence, yet a simple average weighs them equally.
The numbers to put beside the average
- Places listed and places rated. Many places have no rating yet. The average speaks only for the rated ones.
- Reviews per place. A middle value, such as the review count of the typical place, tells you whether ratings rest on real experience.
- The spread. The best and the worst, or how many are rated 4.5 or more, shows whether the market is uniform or split.
- Chains against independents. A chain branch often shares a standard; independents vary more.
Large markets show why the rated share matters. xMap's places database lists 312 hotels within 10 km of the centre of Toronto, and 167 of them have a rating, 3.8 out of 5 on average (November 2026). Nearly half say nothing in that average. In a small town, the same gap can leave the average resting on a handful of places.
A worked example: dentists in a small town
This example is made up. A dental group is deciding between two small towns. Both show an average rating of 4.4 for local practices, so on paper they look equally tough to compete in.
| Town A | Town B | |
|---|---|---|
| Practices listed | 8 | 9 |
| Practices with a rating | 8 | 5 |
| Typical reviews per practice | 140 | 6 |
| Rated 4.5 or more | 6 | 2 |
| Lowest rating | 3.1 | 4.0 |
Made-up figures for illustration only.
Town A's average rests on many reviews: patients like most of its practices, and one weak one pulls the figure down. A newcomer would have to be very good. Town B's average rests on five practices with a few reviews each, and four practices have no rating at all. Its 4.4 means little. The group would learn more there by looking at how long each practice has been open and how busy the high street is.
Misreadings to avoid
- Ranking towns by average alone. Two towns with the same average can be entirely different markets.
- Treating no rating as a bad rating. An unrated place is unknown, often new. Leave it out of the average and count it separately.
- Comparing a town with a city. A city's average is steady; a town's jumps with each new review. Compare like with like.
- Reading one snapshot. In a small market, check again in a month. A single new review can change the picture.
How to do it with an AI employee
An AI employee in xMap Flows can work out the fuller picture for every town you are weighing, every month. The blocks, in order:
- Schedule: the first Monday of each month.
- Area of interest: the towns or suburbs you are comparing.
- Nearby places: the type of place you care about, open places only.
- Calculated field: marks each place as rated or not, and as well reviewed or thinly reviewed.
- xMap LLM: set to work on the whole list, it writes a short note per town on the count, the rated share, the spread and the outliers.
- Email: the note and the list to your team.
What arrives is one note per town, written from the figures it found, with the places sorted from best to worst rated and the thinly reviewed ones marked, so nobody mistakes a five-review 5.0 for a sure thing. The template How customers rate plumbers in my city shows the same idea for one trade.
Questions
How many places make a market large enough to trust its average?
There is no fixed number, but the fewer the places and the fewer the reviews each, the more one place moves the average. Always show the count and the spread beside it.
Should unrated places count as zero?
No. A place with no rating has no score yet. Count it separately and leave it out of the average.
Is a median better than an average here?
A middle value is less moved by one outlier, so it helps in small markets. Showing the share rated 4.5 or more does a similar job in plain words.
Templates that do this
Read next
More on business and places data · Every article · Templates