Year Over Year: What Changed in Listings
Last year listing conventions
The 2025 reading recorded four conventions that held across most of the pages sampled. Measurements were published as an image of a size chart pasted into the gallery rather than as typed text. Batch language sat in the title, usually within the first six words. Photograph sets ran to three or four frames. Video, where it existed, was a link in the description that led off the listing entirely. None of that was unusual; those were simply the conventions most pages happened to share.
Price structure was simpler. Most pages priced two options, and the gap between them was wide enough to read as two different items rather than two tiers of one. Placeholder titles were longer, which made them easier to spot: a string of four or five warm phrases filled the whole headline and could be counted at a glance. A reader deciding whether a page was worth opening could settle it from the headline alone, and that is the convention whose disappearance has cost the most time in the second reading.
Return terms lived on the storefront page rather than on the listing, so a reader arriving from a link list saw the item page without the returns wording attached. The desk recorded that as a structural feature rather than a policy difference, since the terms existed either way; they sat one navigation step away from the place where the decision was being made. Where that wording sits was recorded for every page in the sample, which is what made the move described further down measurable at all.
The 2025 pages were also more uniform in how they presented a specification. Where a material was named, it was usually named in the title and repeated in the description, and that redundancy made comparison cheap: two pages could be placed side by side and read down the same fields. Sampling for that reading was 180 pages across eight categories, read in September 2025 and recorded field by field. The same addresses were re-read in September 2026, which makes the comparison below a matched reading rather than two unrelated surveys, and pages that had disappeared by the second pass were recorded as gone rather than counted as unchanged.
This year shifts
Five conventions moved in the re-read, and four of them moved in a consistent direction. Measurements migrated out of the chart image and into a typed line in the description, where a search engine can index them. Batch language moved out of the title and into the option list, which removed it from the part of the page a search result displays. Photograph sets widened from three or four frames to a fixed sequence of six, which is long enough to cover angles that used to be requested by message.
Video changed place rather than growing. Clips that used to sit behind a description link are now embedded in the gallery, which means the same footage is visible without leaving the page and is also less likely to carry a date. The recorded effect is that a reader now sees movement earlier and sees provenance later, since an embedded clip usually arrives with no visible capture date attached. A clip and a dated still therefore answer two different questions, where previously both sat behind the same link.
Price ladders changed shape. Two-step ladders became three or four steps with narrower gaps, and the lowest step is more often a different build rather than a discount on the same one. Placeholder titles shortened: the same phrases are still in use, but they now occupy the first four words instead of the whole headline, which makes them harder to count and easier to miss.
Return wording moved onto the listing footer, one step closer to the decision point, and that is a change in visibility rather than in terms, since the terms themselves were recorded as broadly stable across both readings. What moved is where a reader has to look, which affects how often the wording is actually read. Not everything changed either: material naming stayed as inconsistent as it was, size charts remained common in the categories where fit varies most, and the share of pages carrying a legible code somewhere in the gallery was unchanged within a few pages. Reporting the stable conventions matters as much as reporting the shifts, because it shows which parts of a listing resist fashion.
Practical impact on judging
Each shift changes where a check has to happen, and two of them change the order of the checks. Batch language in the option list means the label is now read after the page title rather than before it, so a reader working top-down meets the price before the quality word. Measurements in typed text mean a fit check can be run from the description without opening a chart image. Both shifts concern where attention lands first, and neither changes what the check itself looks for.
Embedded video shortens the time to a first impression and lengthens the time to a provenance reading. A clip shows a hinge moving or a zip running, which a still cannot. It does not show when the clip was captured, which a dated still often does. The desk's reading order keeps the dated still ahead of the clip for that reason, and treats the clip as a second look at a part already judged. Where both exist, the pair is stronger than either alone; where only the clip exists, the part is recorded as seen rather than as verified.
Narrower price ladders make the option list more important than the visible figure. Where four steps sit within a small range, the top step is often the same construction as the bottom one, and where four steps span a wide range, they usually are not. The spread, rather than the number of steps, is what tells a reader whether the options are tiers of one item or several different items sharing a page.
Shorter placeholder titles raise the cost of the first check. A long string of warm phrases announces itself; four words inside an otherwise ordinary headline do not. Reading the first four words as a field of their own, rather than reading the headline as a whole, restores the speed the longer convention allowed. Return wording on the listing footer changes what can be assumed at the moment of payment as well: a reader who used to open the storefront page now sees the terms beside the price, which shortens the path and reduces the chance that the wording is read in full. Recording where the wording appears, rather than only what it says, keeps that difference visible in a set of notes.
What to change
Four adjustments follow from the re-read, and each one replaces a step that no longer fits the pages as they are written now. Read the first four words of a title as a separate field. Read the option list before the visible figure, and record the spread as well as the count of steps. Treat an embedded clip as evidence about movement only. Look for return wording on the listing before opening the storefront page.
The recording fields need one addition. A date is no longer enough on its own, because a page can change convention between two readings while keeping its address and its title. Recording which convention the page used, alongside the date, allows a later reading to separate a change of supply from a change of layout, and those two lead to different conclusions about the same page.
Two things should not change. The reading order still runs from what cannot be undone to what can be negotiated, and the evidence level required at each step is unchanged by any of the shifts above. A page that moves its measurements into text has made them easier to index and no more reliable than they were while they sat inside an image. The thresholds themselves are untouched, and so is the rule that a step with insufficient evidence issues no verdict rather than a lenient one.
The next re-read is due with the 2027 archive batch and will use the same fields, with the convention note added. Addresses that have disappeared will again be recorded as gone rather than counted as stable, and any category where more than a fifth of the sample has gone will be reported with that figure attached rather than summarised as a trend. For readers who keep their own notes, the useful unit is the convention rather than the page. A note saying that measurements are typed, batch sits in the options and returns sit on the footer describes any page in the sample well enough to compare against next year's version, while a note that copies a title does not.
Listing conventions that moved between the 2025 and 2026 readings
Matched re-reading. The desk re-read 180 public listing pages in September 2026 that it had first recorded in September 2025, using identical fields on both passes across eight categories. Only conventions that changed on at least a quarter of the re-read pages are listed; the rest are recorded as unchanged and left out. The comparison covers page wording and structure and says nothing about what any page delivered. Addresses that had disappeared were excluded from the convention counts. Last checked 2026-09-26.
| Item | Reading |
|---|---|
| Measurement statements | Moved from a size chart image into a typed line in the description |
| Batch labels | Moved from the title into the option list |
| Photograph sets | Widened from three or four frames to a fixed six-frame sequence |
| Video | Moved from a description link to a clip embedded in the gallery |
| Price ladders | Moved from two steps to three or four, with narrower gaps |
| Placeholder titles | Shortened; the same phrases now sit inside the first four words |
| Return wording | Moved from the storefront page onto the listing footer |
Sources and checks
- Public community thread — Public listing pages recorded in September 2025 and re-read in September 2026 with identical fieldsn=180
- Editor analysis — The convention list and the quarter-of-sample threshold used to decide what counts as a moveno sample
Last checked · next review due · status current.