Finding things in an archive

Where the final brand video is saved, and how to find it

Four places hold it, almost always: the thread where it got approved, the delivery link the editor or agency sent, the platform it was published on, and a shared drive folder under a filename that made sense to one person during one week. Start with the published copy, because it is the only version you can be sure someone signed off on, and the name it carries there is usually the best search term you have for finding the master.

Start from the copy that went live

The file that ran is the file that cleared review. Pull it from wherever it was published: the ad account, the video platform, the page it sits on. What you get is a compressed export rather than a master, which is fine if you need to show it to someone and a problem if you need to re-cut it. Write down the exact name and the date it went up before you go looking for anything else.

Search the approval thread, not the drive

Files get moved. Conversations about files stay put. Search chat and email for the words people used at delivery time: "final", version numbers, the campaign name, the words "here's the link". A two-year-old message often still resolves to the folder the editor actually used, which beats guessing at folder names. This is also how you find out that the thing you remember as one video was three cuts, and which one the client approved.

If the folder is the problem, naming is the fix

When this happens twice in a quarter, better searching is not the answer. A naming convention plus one canonical location costs nothing but discipline, and it protects only the files created after you adopt it. A shared library with permissions and review states buys you one approved version per asset and a place where "final" means something, which is the part an asset process for video is really paying for; the cost is that somebody has to own it, and old files do not migrate themselves. A production-side system that keeps masters alongside proxies solves the other version of this, where the export is fine but the source is gone. Keyword search in any of them finds only what a person typed in, so the payoff tracks how honest your team is about filling fields.

When you need the footage, not the export

"Where is the final video" and "where is the shot of the customer describing the rollout" are different questions with different answers, and the second one is not a filing problem. One route is to put footage into a layer that indexes what is inside each video once at upload, then query it by description instead of by filename. What comes back depends on the search mode: a quick pass tends to tell you which videos are probably relevant, and narrowing to the seconds inside them takes a slower, more precise pass that you wait on. Vivu is one implementation, so "which of last year's shoots has the founder explaining the pricing change" gets asked inside a project instead of answered by opening folders, though it will not tell you which export the client approved, because that fact only ever existed in a thread.

When none of this is worth doing

If your team finishes a handful of videos a year and one person exports all of them, a naming convention and a single folder will outperform anything you buy. Same if the file you are missing right now is one deliverable and the editor still has the project open. Ask them, then write down where the answer was. Systems start earning their keep when the number of people who need to find things is bigger than the number of people who made them.

Which problem you actually have

If one person knows where the file is and it was never written down, this is a process problem, and the cheapest fix is deciding that the published version is canonical and naming it so. If nobody can answer because nobody can describe what is inside hundreds of files, this is a retrieval problem, and it gets worse at the rate you shoot. Ask what happened the last three times: if the same person always eventually finds it, you need a habit, and if the searching itself was the hard part, you need something that reads the footage. Tracking down a video you made years ago is the same exercise with a colder trail.

FAQ

What do I do if the editor who made our final video has left the company?

Start with the published version and the delivery link, both of which live outside their accounts. Company chat and email history usually survive a departure, so search those for the handoff message rather than trying to get into their drive. If the master is genuinely only in their personal storage or in an agency's project, the practical move is to ask, in writing, for the master and the project file, and to treat the compressed published copy as what you have until it arrives.

Is the version on YouTube good enough to reuse?

It is good enough to watch, reference, and link. It is usually not good enough to re-cut or to run as a paid ad, because it has been compressed twice and often has burned-in end cards or captions. Use it to confirm you have identified the right video, then decide whether the reuse you have in mind survives that quality, and go looking for the master only when it does not.

How do I keep this from happening with the next campaign?

Decide before delivery where the final file lives and what it is called, and put that in the brief rather than in the handoff. Two rules cover most of it: one folder per campaign that holds the approved export, and a filename with the date and the campaign in it instead of a version number. Version numbers are what produce the "final_final_v3" problem, because they record the editing process rather than the decision.

Can I search inside old video files to find the right one?

Yes, if something has indexed what is inside them. Filename and folder search only knows what a person typed, so it cannot find a video by what happens in it. The tools that can either read a transcript, which works when the thing you want was said out loud, or index the picture and sound so you can describe a shot. Both need a copy of the video to process first, which is the step people usually discover late.