Titles, descriptions and metadata
Decided alongside the edit and never bolted on afterwards, and pointed at the one destination the funnel actually needs.
The short version
Is this for you? For titles bolted on at upload and descriptions that point three ways at once. It pays most where the description is doing real funnel work; if there is nothing to send a viewer to yet, we will say the lever is small.
What leaves your plate. Metadata written by the people who just cut the video, while they still know which forty seconds are the reason to watch, and pointed at one destination rather than three.
What starts working. A page that reads as if one person built all of it: a clear promise, a video a learner can navigate, and one obvious thing to do at the end.
The smallest way in. A free audit: three fields, and a written reading back, usually within about three working days. It commits you to nothing.
What to read next. The five levels to see where this fits, or the casework to see it done.
Metadata written while somebody still remembers the video
Metadata is written while somebody still remembers the video
Titles, description and chapters are drafted by people who have just spent hours inside the piece and know which forty seconds of it are the reason to watch. Metadata written days later, from a filename, by somebody who never saw the cut, is a guess dressed as optimisation. The difference shows up in the first line of every description.
One destination, and the page is built around it
A description that offers four routes has chosen none of them. The invitation points at a single place the funnel actually needs, the timestamps make the video navigable instead of decorative, and everything else on the page is arranged to keep that one route legible on a phone with the description collapsed.
The research is into words, not into tags
What a niche searches for is not what its adjacent niche searches for, and a term that carried one brand will quietly underperform for another, so the research runs per category and per channel: the phrases with real demand, and the words the audience actually uses rather than the ones the industry does. Those words earn their keep in the title, the first lines of the description and the spoken track, because that is where the platform reads them. The tag field itself moves very little on its own, and we treat it as such: filled lightly, mostly with the spelling variants and near-misses a viewer might actually type, weighed rather than stuffed, and never sold to you as the thing that made a video work. The platform's own guidance says as much: the title, thumbnail and description are the metadata that matter for a video's discovery, and tags mainly help when the subject is one people commonly misspell.
Subtitles are part of this job
Caption files are prepared as standard and routed only once the cut is approved, so subtitles never chase a moving edit. They serve accessibility first and search second, and both of those are real. Where a partner publishes for themselves, the files arrive in the video's own folder with the music licences beside them.
Preparing the page a long-form video lands on
What is prepared for each long-form piece
Several title options, so there is something to choose between. A description written for a reader and indexed by a machine. Chapters and timestamps. Playlist placement. Tags filled from that same research, a minor field kept honest rather than stuffed. Caption files in SRT and plain text, and the end screens and cards that decide where a finished viewer goes next.
How it is kept honest
Every one of these choices is checked later against what the channel actually did: whether the metadata, the playlists or the packaging moved click-through and watch time, or whether something else did. Optimisation that is never measured is a preference, and preferences accumulate in a channel until nobody remembers which of them was ever tested.
Metadata written after the export is a guess about a video everybody involved has already stopped thinking about.
It is written while somebody still knows which forty seconds are the reason to watch.
What changes for you
A page that reads as though one person built all of it, which is roughly what search rewards and exactly what a viewer needs: a clear promise, a navigable video, and one obvious thing to do at the end of it.
Related: Thumbnails · Between the uploads · CTA and funnel design · Subtitles and captions
The audit reads the channel the way a manager would: cadence, packaging and what the analytics are asking for next. What we would change comes back in writing, before any commitment.
Request your free audit.
Three fields, and a written reading back: a fit snapshot, two or three prioritised opportunities, and a recommended next step.
A person reads the channel and writes the audit by hand: a considered read typically takes three working days. That is the usual shape, not a promised turnaround. We use these details only to reply to you: no lists, no lurking.
What you will get
A fit snapshot: where your channel stands, and whether we are a match.
Two to three opportunities: specific, prioritised, yours to keep.
A recommended next step, even if that step is not us.
The audit is free and commits you to nothing: nobody follows up with a call you did not ask for.
The casework shows all of this done.
The case studies are the same work with the numbers, the sources and the dates attached.
See the casework →