A product can hold a respectable score from professional reviewers and a polarised average from owners. The gap is structural rather than a sign that one side is wrong.
Owner ratings are self-selected
Buyers who leave ratings are not a random sample of buyers. People who had a problem and people who are delighted both have a reason to write; the merely satisfied usually do not.
This produces the familiar shape where ratings pile up at five stars and one star with little in between. The average sits somewhere neither group intended.
The distribution carries more information than the mean. A product with a heavy tail of one-star ratings has a failure mode worth reading about, even if the average looks acceptable.
Reviewers compare, owners experience
A reviewer has handled competing products recently and rates against that set. An owner rates against expectation, price paid and whatever the previous device did.
Those are different questions. A phone that is unremarkable next to its rivals can still be a large improvement for someone upgrading after four years.
Neither framing is more valid. They simply answer different things, and combining them into one number discards the part that made each useful.
Delivery and support enter owner scores
Owner ratings absorb everything about the purchase, including shipping damage, a missing accessory, a slow refund or an unhelpful support call.
Reviewers rarely encounter any of it. Their unit arrives directly from the manufacturer, and their support queries reach a different queue from an ordinary customer's.
A large share of low owner ratings describe the transaction rather than the product. Reading the text, not the star count, separates the two.
Timing changes the picture
Early owner ratings come from enthusiasts who wanted the product enough to buy it immediately. Later ratings come from broader buyers with less tolerance for rough edges.
Failures also arrive late. A component that dies after a year appears in ratings long after launch, by which point the review cycle has moved elsewhere.
Sorting owner ratings by date rather than by helpfulness often reveals a trend that the headline average has flattened out.
Scores compress in both directions
Professional scores cluster because outright bad products are rare among devices worth reviewing, and because a very low score invites disputes over methodology.
Owner scores spread because there is no methodology to defend and no relationship to protect. Each system distorts in its own direction.
Treating the two as independent readings, rather than as competing estimates of one true value, makes both more useful than either alone.