The five spreadsheet formats and what each one hides

A format is not a skin on the same content. It decides what can be sorted, what can be compared across entries, and what disappears because there is nowhere to put it. The conclusion, before the detail: a flat sheet is best for bulk scanning and worst for currency; a faceted list is best for filtering and worst for completeness; a card grid is best for a first impression and worst for comparison; a video-first list is best for showing an item in motion and worst for search; a mixed index is best for range and worst for consistency.

The reason this matters more than it sounds is that most readers choose a list by how large it looks rather than by how its format constrains it. A format that cannot show a review date will not show one, no matter how many entries it holds, and a format that cannot show two entries side by side will never let you compare them without opening both.

So the useful question is not which list is biggest. It is which format supports the decision you are about to make, because a mismatch between format and decision is what produces the feeling that a list is huge and useless at the same time.

A flat sheet is a grid of rows and columns. It wins on scanning density: you can see thirty entries at once, sort by a column, and hold a whole category in your head. For a bulk order of unbranded basics, where the decision is about price band and availability rather than about any single item, nothing else comes close.

What it hides is context. A grid has no room for a paragraph, so the reason an entry was included cannot be written down. It also tends to hide the shape of its own data: a column filled from one source and a column filled from another look identical, and a column that is empty in a third of rows is easy to miss when scrolling.

The practical test for a flat sheet is whether the column headers tell you the source of each column, not just its name. A header that says "updated" is ambiguous; a header that says "price last checked" is not. Where the headers are ambiguous, treat the sheet as a starting list rather than as a record.

A faceted list is a set of entries with filters on top: by category, by marketplace, by size range, by season. It wins on narrowing. Where a flat sheet asks you to scroll, a faceted list asks you to choose, and choosing is faster when you already know what you want.

What it hides is whatever falls outside the facets. A list organised by category cannot show an item that belongs to two categories comfortably, and most items belong to two. A list organised by marketplace hides the fact that the same product is listed on three marketplaces at different terms. The facets are not a neutral filter; they are a set of decisions about what counts as a dimension.

The practical test is to count the facets and then ask what is missing from them. A list with four facets has chosen four dimensions and excluded everything else. If the dimension you care about — dispatch speed, return terms, declared value band — is not among them, the format cannot help you find it, and you will end up checking entries one at a time.

A card grid gives each entry a box with an image and a couple of lines. It wins on first impressions and on browsing: you can tell at a glance whether a list is in the right territory, and the visual rhythm makes a large set feel navigable rather than daunting.

What it hides is comparability. Two cards cannot be placed side by side meaningfully, because the fields inside them are not aligned. Sorting is usually impossible, and searching inside the list is often absent. The format optimises for the moment of discovery and offers nothing for the moment of decision.

The practical test is whether the card carries any field you could sort on if you wanted to. Where a card shows a title, an image and nothing else, the list is a browsing surface and not a working tool. That is a legitimate thing to be, as long as you do not try to use it for comparison.

A video-first list leads with short clips. It wins on showing behaviour that a photograph cannot: how a fabric moves, how a zip runs, how a sole flexes. For an item where the material is the risk, a few seconds of footage carries more information than any still image.

What it hides is everything searchable. A clip has no text to index unless the maintainer writes it, and the writing is usually a caption rather than a field. There is no review date, no link shape, no duplication figure, and no way to check whether the same item appeared in an earlier version of the list.

The practical test is whether the list also carries a text record. Where the clips are the only artefact, the list is not checkable and its entries cannot be compared with anything. Where a text record exists alongside the clips, the format is genuinely additive rather than a replacement.

A mixed index combines formats deliberately: a faceted header, a flat table for the bulk, cards for a featured subset, and links to longer write-ups. It wins on range, because each part of the format is chosen for the job it does best.

What it hides is consistency. Where different sections are filled from different practices, the reader cannot tell whether an empty field means absent or not applicable. Mixed formats also tend to have mixed review dates, so an entry in the featured strip can be far older than an entry in the table below it, with nothing on the page saying so.

The practical test is whether the page states which part of it was reviewed when. A mixed index that carries one date for the whole page is telling you that parts of it are older than that date and that the maintainer did not distinguish them.

Two cases break the pattern. Where a list is small, under a hundred entries, the format barely matters, because you can read all of it and the constraints never bind. And where a list is maintained for a single person, its format reflects one workflow rather than a general one, which makes it excellent for that person and unpredictable for anyone else.

A third case is worth adding, because it is common and it defeats the test above. Where a list changes format between visits, which happens when a maintainer migrates a sheet into an app or rebuilds a page, the fields you relied on may survive in name and change in meaning. A column called condition that once meant new or used can come to mean a seller grade, and nothing on the page announces the change.

The defence against that is the same as the defence against a stale review date: keep the date you recorded, not just the field you recorded. A field with a date attached can be re-interpreted later; a field without one cannot.

Neither exception is an argument against checking the format. It is an argument for checking the size and the audience first: a small list removes the format question, and a personal list answers it in a way that may not transfer.

Four checks, in order, and each one takes about thirty seconds. First, find the date and see whether it covers the whole page or a section. Second, look at the headers or the card fields and ask what source each one came from. Third, try to sort or filter on something you care about and see whether the format allows it.

Fourth, take two entries from different parts of the list and answer one question about each from the record alone. If the record cannot answer a question you would ask before ordering, the format has excluded that field, and no amount of scrolling will bring it back.

Where a format passes all four, it is worth using regardless of how large it looks. Where it fails the first, treat every entry as undated; where it fails the fourth, treat the list as a browsing surface and plan to verify each candidate individually.