There are two ways people arrive at a catalogue, and they want opposite things. One knows what they're looking for. The other doesn't, and is hoping the interface will help. Most sites build for the first and leave the second to scroll.
The two behaviours
Searching is target-directed. You have a title, an artist, or a specific element in mind, and success means finding it in one step. Ranking quality matters, spelling tolerance matters, and everything else is noise.
Browsing is exploratory. You have a rough mood and no target, and success means encountering something you wouldn't have thought to ask for. Ranking barely matters; what matters is coverage, variety, and the ability to rule things out quickly.
These need different interfaces, and merging them into one search box serves the first adequately and the second not at all.
What browsing actually requires
Four things, none of which are a search box.
- Meaningful sorting. Newest, most-viewed, and random each surface different material.
- Exclusion filters. Ruling things out narrows a huge catalogue faster than including does.
- Visible facets. Artist, language, format, and length shown as filters rather than buried.
- A sense of position. Some indication of how much you've seen and how much is left.
The last one is the rarest and the most valuable. Infinite scroll deliberately removes any sense of progress, which suits an advertising business and frustrates anyone trying to actually look through something.
A history of what you have already opened helps almost as much. Without it, a long session becomes a loop, and the same handful of entries keep resurfacing while everything below them stays unseen. Marking visited entries is a small change with an outsized effect on how large a catalogue feels.
Why exclusion beats inclusion
In a catalogue of a hundred thousand entries, adding a filter that includes a tag still leaves thousands of results. Removing four tags you don't want cuts the field far more usefully.
| Approach | Effect on a large catalogue |
|---|---|
| Include one tag | Still thousands of results |
| Include three tags | Often too few, misses near matches |
| Exclude several tags | Removes most of what you would skip |
| Sort randomly, then exclude | Best coverage for exploration |
That's why exclusion is the main use of any tag system, and why inconsistent tagging breaks browsing before it breaks search — the failure modes described in our piece on community tagging.
What search needs that browsing doesn't
Spelling tolerance, alias handling, and the ability to search by artist as a field rather than as free text. That last one is the biggest gap in most catalogues: a name typed into a general box competes with every entry that mentions it.
Namespaced fields fix it, which is one more argument for proper metadata rather than a flat tag list — the six fields we set out in the piece on metadata.
Alias handling matters because the vocabulary in this field is fragmented across spellings, as covered in our piece on the spelling split. A search that treats variants as unrelated returns four partial result sets.
Language handling is the third gap. Titles exist in an original form, a romanisation, and one or more translations, and a search that only indexes one of them will miss the work entirely for most people who look for it.
The things neither mode should compromise
Verified age entry loads before any content, whichever way someone arrived. Artist credit appears on every entry. Exclusion filters apply to recommendations as well as to results, since a filter that only works on one screen isn't a filter.
And the floor holds regardless: characters are fictional and unmistakably adult, with no ambiguity about age, no real-person sexual depiction, and no coercion presented as romance. Our standards are in the control room, and connected creation with real account and moderation controls stays on models.
Designing for people who don't know what they want
Most sessions start without a target. That's the fact most catalogue interfaces are built to ignore, and it's why so many large collections feel unusable despite containing exactly what someone wanted.
Give people a way to rule things out, a way to sort, and a sense of where they are. Those three additions turn a wall of thumbnails into something you can actually explore, and none of them require a better search algorithm.