Overtones 04Data study

Five Stars for Everything.

Almost every store listing for headphones, IEMs and speakers leads with its star rating. Those stars say little about what reviewers concluded. Owners’ favourites often land near the bottom of the reviewers’ list, partly because the two are judging different things.

171 Overreviews analysed
The data
13 of 25of owners’ favourites land in the reviewers’ bottom quarter
171 Overreviews · 4,419 human reviews

The Meze Audio 109 Pro and the Sivga SV021 Robin are both headphones, and owners give both 4.3 stars. Reviewers don’t see them as equals: they rank the first near the top of the catalogue and the second near the bottom. Nothing on a store page would tell you that.

That turns out to be the rule rather than the exception. Every Overreview on Audiowords records the owner rating where one exists, almost always Amazon’s, next to a consensus score built from what reviewers actually wrote. This piece sticks to headphones, IEMs and speakers, where a verdict is mostly a matter of opinion. In all, 171 of them have a solid owner rating, from at least 20 owners. That’s enough to ask a simple question: do products with more stars get better reviews?

If they did, the chart below would show a cloud of dots climbing from left to right. It doesn’t. If anything, it tilts slightly the wrong way.

Figure 1Live data

Owner stars and reviewer scores

One dot per product with at least 20 owner ratings. Named: the products owners and reviewers rank furthest apart.
Live data from 171 headphones, IEMs and speakers with owner ratings, updated 25 September 2026
Image

Here’s a plainer test. Take the 25 products owners rate highest. Reviewers put only two of them in their own top 25: the Neumann KH 120 II and the Sennheiser HD 600. Picking at random would manage about four. Follow all 25 into the reviewers’ ranking, and more than half end up in its bottom quarter.

Figure 2Live data

Where owners’ top 25 land with reviewers

Each line runs from one of owners’ 25 highest-rated products to its place in the reviewers’ ranking of all 171 headphones, IEMs and speakers.

Reviewers’ top quarter: 2In between: 10Reviewers’ bottom quarter: 13
Live data from 171 headphones, IEMs and speakers with owner ratings, updated 25 September 2026
Image

That bottom quarter includes some of the best-known names in audio. The Sony MDR-7506, the Sony WH-1000XM3 and the Beyerdynamic DT 990 PRO (250 Ω) all end there, each with tens of thousands of owner ratings behind it.

Everything is four and a half stars

Part of the problem is the scale itself. Owners give nearly every product here four stars or more, and half the catalogue sits between 4.2 and 4.6. That leaves a few tenths of a star to separate a favourite from an also-ran.

The reviewer score is built to leave that room. It runs from 0 to 100 and puts a mixed verdict at 60, near the middle, so a middling product looks middling. Clear praise scores around 80, and anything under 60 leans critical. It also moves in whole points instead of tenths of a star, so smaller differences still show.

The two headphones at the start of this piece make the point. Both show 4.3 stars on a store page. The reviewer score gives the Meze Audio 109 Pro 80, “well liked”, and the Sivga SV021 Robin 58, “leans critical”.

Where they part ways

The biggest disagreements run in both directions. Owners put the Audio-Technica ATH-M50x near the top of their list, with tens of thousands of ratings behind it. Reviewers put it second to last of all 171. One reviewer warns up front that “this is NOT a neutral-sounding headphone”1.

The Yamaha HS5, a studio monitor, is another owner favourite that reviewers rank near the bottom. Audio Science Review measured it and put the low end bluntly: “Alas, it has no bass or low end.”2

Going the other way, reviewers put the KEFINE Delci near the top of their list. Owners leave it near the bottom of theirs. One reviewer calls it “the Toyota Camry of the IEM world, not super exciting”3. It’s meant as praise, but it may explain why buyers are less keen.

Sound is only part of it

Much of the gap probably comes down to what each side is judging. Critical reviewers mostly judge the sound: its balance, its detail, how it stacks up against its rivals. Owners judge everything they live with. Is it comfortable after a few hours? Does it still work after a year? Does it look good? Was it worth the money? That goes double for wireless earbuds and designer headphones, where sound is only part of what you pay for.

None of this makes owner ratings useless. They answer a different question: were buyers happy with what they chose? So before you trust either number, ask what you want from the experience. If sound comes first, read the reviewers. If comfort, looks or value come first, owners may tell you more. Where an owner rating exists, each Overreview shows both side by side.

The charts redraw whenever a new Overreview is published. This version covers 171 headphones, IEMs and speakers with owner ratings, as of 25 September 2026.

Data & method.

171 Overreviews · 4,419 human reviews
  • What was counted171 headphones, IEMs and speakers whose Overreview records a retail owner rating built from at least 20 ratings, out of 205 that record one at all. Nearly all of those aggregates are Amazon’s. Rankings and percentile ranks are within these 171 products.
  • Why only headphones, IEMs and speakersThey’re where a verdict is mostly a matter of opinion. Amps, DACs and controllers are judged far more on function, and their reviewer scores run higher across the board, so ranking them alongside would measure product type as much as owners against reviewers.
  • The reviewer scoreThe 0–100 consensus score is built from each Overreview’s topic verdicts, weighted by what matters for the product type. “How review scores work”, linked from every review, has the details. A neutral verdict scores 60. For headphones and IEMs, most of the weight is on sound and the rest on comfort, build, value and isolation. It summarizes human reviewers, not owners, and it is not an average of other sites’ scores.
  • Order, not levelsOwners use one to five stars, where one means bad; the reviewer score puts a neutral verdict at 60 out of 100. Any conversion between the two is a choice that moves the apparent gap, so the study compares order instead: the correlation between stars and score (r = -0.15), the rank correlation (ρ = -0.18), each product’s rank on both scales, and the overlap of the two top-25 lists, where chance alone would give about 3.7. Ties go to the product with more owner ratings, or with more reviews behind its score. None of these change under a conversion.
  • CaveatsOwner aggregates are self-selected and skew high, rating counts range from a few dozen to thousands, and a store listing can pool several variants. Named examples need at least 50 ratings. The catalogue leans toward gear enthusiasts discuss, not the whole market.
  • Live dataThe charts and every number in the text are recomputed from the records each time the site is built. The CSV holds one row per product: owner stars, rating count, consensus score and both percentile ranks.
Download the data (CSV)

Reviews quoted

  1. 1
    Audio-Technica ATH-M50x

    Overreview 16 sources

  2. 2
    Yamaha HS5

    Overreview 14 sources

  3. 3
    KEFINE Delci

    Overreview 32 sources