What changes when you have 200 videos

At twenty videos you can hold the whole library in your head. At two hundred you cannot, and everything you were doing informally has to become a system — usually at the worst possible moment, when a product changes and nobody can find which videos show it.

Naming stops being optional

At twenty, “blue shirt one” is fine. At two hundred it is unsearchable.

A convention that survives: product, then variant, then what the video is. “Linen shirt — sand — fit and drape”. Predictable, searchable, and it tells you what the video is without opening it.

Retrofitting this across two hundred videos is a genuinely miserable afternoon. Starting it at thirty costs nothing.

Playlists become the structure

One playlist per collection you actively merchandise, one per campaign, one for the home page. A video can be in several.

The anti-pattern is a playlist per placement — “home page top”, “home page bottom” — which ties your content structure to your layout, so a layout change becomes a content migration.

Tagging has to happen at import

At small scale, untagged videos get tagged eventually. At large scale, “eventually” becomes a backlog of sixty untagged videos that nobody will ever work through, and every one of them is a video earning nothing.

The rule that holds: a video is not in the library until it is tagged. Untagged means still in progress.

Pruning becomes the main job

This is the counter-intuitive one. Past a hundred videos, removing them matters more than adding them.

Discontinued products, dead campaigns, three near-identical videos of the same thing. Each one occupies a placement, dilutes the average quality of what a shopper taps, and makes the library harder to work in.

Once a quarter, remove ten per cent. It is always possible and it always improves things.

Audit what the catalogue did to the videos

Products get renamed, variants get retired, prices change, things get unpublished. Videos tagged to them do not update themselves in the sense that matters — the tag may still resolve while pointing at something that is no longer the right answer.

A quarterly check on your top twenty catches the ones that matter. The long tail can wait until it is being used.

Fewer people should be able to publish

At scale, the failure is inconsistency: three people tagging differently, naming differently, placing differently. One person owning the library and a clear convention is worth more than an extra pair of hands adding to it.