How to Get More Views From Old YouTube Videos

Most channels have a shape like this: the last ten videos got real attention, and everything before that got whatever the creator had energy for at the time.

Those older videos don't stop working. They keep surfacing in search, in suggested, in the sidebar of the newer ones. They accumulate impressions for years. And they're doing it with titles written in a hurry in 2023 and thumbnails that were the first frame that wasn't blurry. Our free thumbnail downloader is a quick way to pull your old ones and look at them next to each other, which is usually enough to decide the question.

The gap between what a back catalogue earns and what it could earn is usually the largest single unclaimed thing on a channel. Almost nobody claims it, and the reason is arithmetic.

Why nobody does this by hand

Say a proper pass on one old video takes 40 minutes: pull the transcript, write a better title, rewrite the description in your current template, add chapters it never had, find a decent frame, get it to spec, upload it, note what you changed.

Sixty videos is forty hours. That's a working week doing nothing else, on content you already published, with no new upload at the end of it.

So it never happens. Not because creators don't know the back catalogue matters — everyone knows — but because forty hours is a number you look at once and then stop thinking about.

With ChatGPT in the loop you might get it to 25 minutes a video, which is twenty-five hours. Still a week. The chat window makes each individual step faster and doesn't change the fact that there are sixty of them, because each one is its own conversation that you drive by hand.

What changes when it's a batch

In Growati this is one job: sync the channel, select the videos, choose a template, review the queue as results come in.

The reviewing is still yours and still takes real time — but reviewing sixty generated titles is an afternoon, not a week, because you're reading rather than producing. That's the actual shift. The generation stops being the bottleneck and the bottleneck becomes your judgement, which is where you wanted it.

Which videos to redo first

Not the most-viewed ones. The instinct is to start there, and it's wrong.

Sort by impressions with weak click-through instead. That pairing only exists inside YouTube Studio, since impressions are private; our free video statistics tool covers the public numbers if you want to scan a channel quickly. A video with 80,000 impressions and a 2.1% CTR is telling you that YouTube is already offering it to people and they're declining. That's a title and thumbnail problem with a large audience waiting behind it. A video with 400,000 views and a 9% CTR isn't asking for help.

Second priority: videos with no chapters at all, especially anything over ten minutes. Adding chapters to a long video that never had them is one of the few metadata changes that's almost purely upside. Our free timestamp link generator handles the formatting, and read the chapter rules first — a list that breaks any one of three requirements renders as nothing at all.

Third: anything where your description template has changed since. Old videos pointing at links that have moved are quietly leaking the traffic you did earn.

The risk nobody mentions

Changing the title of a video that is currently ranking well is a real risk, and it's the part of back-catalogue optimisation that gets glossed over in posts like this one.

If a video has settled into a stable position in search for a query, its title is part of why. Rewriting it can improve CTR and can also move the video off the query that was feeding it. We've seen both. There is no reliable way to know in advance which one you'll get.

The conservative approach is to leave your top performers alone and work on the middle of the distribution — the videos with impressions but poor conversion, where there's little to lose. The aggressive approach is to change everything and read the analytics in six weeks. We'd suggest the first, but people who've done the second have sometimes done very well out of it, and we'd be overstating our confidence if we pretended the answer was settled.

Reading the results without fooling yourself

Whatever you change, don't evaluate it against the previous four weeks. Seasonality, an unrelated video taking off, or one old upload getting picked up by the algorithm will all move your numbers more than a metadata pass will.

Compare each changed video against itself, over a comparable window, and expect the signal to be noisy for the first month. If you want the longer version of which numbers are worth watching, we wrote about that separately.

Start with the impressions-with-weak-CTR list. It's usually about a dozen videos, and it's where the whole return is concentrated.