The average rating is the least interesting number
Everyone looks at the star rating. It is the one number that tells you almost nothing, because it is an average of thousands of people who mostly agree, and averages hide the disagreement that matters.
Loop Earplugs sits at 4.35 across 48 products showing reviews, and their best-reviewed product alone carries 18,976 reviews. You cannot read that. Nobody can. And skimming the first page tells you only what the algorithm decided to show you today.
The analysis is not reading them. It is grouping them, so that the recurring problems separate from the one-off ones.
Worked on Loop Earplugs
Their complaints, grouped by theme and counted as a share of all complaints:
Fit and seal: 10%. One in ten complaints is the product not sitting right in the ear.
Not enough blocking: 10%. Another one in ten bought the product for a job it did not do well enough.
Comfort and pain: 9%. Close behind, and distinct from fit: these are people for whom it sat correctly and still hurt.
Price and value: 2%. Almost nobody.
Now look at what that pattern says. Roughly three in ten complaints are about the physical experience of wearing the thing, and one in fifty is about the price. This is not a company with a pricing problem. It is a company whose product fits some ears better than others, and whose buyers are willing to pay but not willing to be uncomfortable.
What that is worth to a competitor
It is ad copy that has been pre-tested by their customers. If a tenth of their complaints are about seal, and your product seals better, that is not a claim you have to invent. It is an objection their own buyers wrote down, and you can answer it in their words.
It tells you which comparison to pick. The useful head-to-head is not your best product against their best product. It is your product against the one at the bottom of their catalogue, ranked worst first, where their reputation is currently being spent.
It tells you what not to compete on. Two percent price complaints is a clear instruction: do not open with cheaper. You would be answering a question their buyers are not asking.
And the shares move. A theme going from 4% to 10% over a quarter is a regression, a new supplier, or a new product that does not work as well as the old one. That is the moment worth catching, and it is only visible if somebody counted the same way last quarter.
The trap in doing this by hand
Reading twenty reviews and forming an impression feels like research and is closer to superstition. The reviews you happen to read are the ones the platform sorted to the top, which are usually the most helpful or the most recent, and neither is representative.
Worse, an impression cannot be compared to last quarter's impression. You will remember that people complained about fit. You will not remember whether it was one in twenty or one in eight, and the difference between those two is the entire finding.
Where these numbers come from
Every review we can reach on the competitor's own store and on the public review platforms, grouped into themes and counted, then held against the same count from the previous collection. The theme shares sit on the Benchmark, and the products behind the rating, weakest first, sit on Details with a link to each product page so you can read the actual sentences.
A percentage you cannot open is a claim. A percentage that opens into the reviews it was counted from is evidence.
Pick one competitor. See what we find.
Your first benchmark is free. One domain, measured against you, no card and no call.