Metadata advice for Shorts is mostly people who rank videos for a living explaining fields that barely move. We went looking for what each field plausibly does before publishing twenty of them, and the honest summary is that the ordering matters more than any individual tip.

One caveat first, because it applies to everything below: the sources for this are SEO write-ups rather than documentation from YouTube. Treat it as direction, not specification, and believe your own channel over any of it.

The ordering

1. The first one to three seconds. Not metadata at all, and it outranks everything that is. Shorts distribution runs on whether people keep watching and replay, and no title compensates for an opening that takes four seconds to reach the point. If you are going to spend effort somewhere, spend it on the first line of the script.

2. The title. Matters less on Shorts than on long videos, where it drives the click. In the Shorts feed there is barely a click to drive. It still does real work in three places: YouTube search, Google's Shorts shelf, and your own channel grid, where people who liked one video decide whether to watch another.

Keep it under roughly 70 characters so it does not truncate on mobile, and put the interesting word in the first two or three. "My Roommate Started Labeling Her Food With Warnings" survives truncation. "A Story About Something That Happened With My Roommate And Her Food" does not, and gives nothing away in the visible half.

3. Hashtags. Three to five, not fifteen. One or two general, one or two specific to the subgenre. The first three from the description appear as a clickable row above the title, which is their main visible job. More tags do not mean more reach; they mean a messier description.

4. Backend tags in Studio. Weaker than the title or hashtags, and cheap enough to bother with anyway. Three or four is plenty: one close variant of your title, one or two adjacent phrases, one broad category term. There is no benefit to filling the 500-character limit and nobody who tells you otherwise has data.

The hashtag we will not use

The obvious tag for our format is #redditstories, or #aita. The genre is recognisable through those tags and they carry real traffic.

We do not use them, because our scripts are original. Nobody posted these stories on Reddit; we wrote them. A tag that implies a source which does not exist misleads people about where the content came from, which is a different thing from a tag being merely aspirational.

So the tags we use describe the genre without claiming a platform: #pettystories, #storytime, #relatable. If you write your own stories, the same applies. If you genuinely adapt real posts, then the tag is accurate and you have a different question to answer, which is whether you have the right to adapt them at all.

What we could not verify

Whether any of this moved anything. Across twenty videos our titles and tags were consistent in style, and the variation in views was enormous: 11 at the low end, 1,269 at the high. Whatever produced that spread, it was not the metadata, because the metadata barely varied.

That is worth saying plainly because most articles on this topic imply a causal link they have not tested. Ours would have been in a position to test it and did not, because we were testing something else.

The practical version

Write the first line of the script until it earns the next three seconds. Write a title that survives being cut in half. Pick four hashtags, two general and two specific. Put four words in the backend tags. Then stop and go write the next script, because that is where the difference actually comes from.

If the writing is the part that slows you down, we have a piece on what a script for this format actually requires. If the assembly is, that is the part Memecut does: script in, voiceover, captions and a reaction meme on each line out.

Sources: YouTube hashtags guide, Shorts hashtags, title versus description, Shorts SEO guide.