Every December I open a few analytics dashboards and find the same thing. Somewhere between 60 and 80 percent of a channel's views that month come from videos published in a previous year. The creator is exhausted, shipping one more upload before the break, and completely ignoring the library that is quietly doing most of the work.
That is the whole argument for an end of year back catalogue refresh. You already paid for those videos. The research, the filming, the edit, the upload. The only thing standing between an old video and a second life is usually the packaging, and packaging is the cheapest part of the stack to redo.
I run this pass with clients every year between late November and mid January. Here is the actual process, including the parts where it does not work.
What is a back catalogue refresh, exactly?
A back catalogue refresh is redesigning the thumbnail and title of videos you already published, so they compete properly in browse, suggested, and search today. You are not re editing the video. You are not reuploading. You are changing the promise on the outside of the box.
The modern name for that promise is packaging. Thumbnail, title, and the first frame the viewer sees when the video starts, designed as one idea rather than three separate tasks. Thumbnail design on its own is the old framing, and it is why so many refreshes underperform. People swap the image, leave a title written in 2023, and then conclude that repackaging does not work.
The video did not fail. The sentence above it failed, and nobody rewrote the sentence.
Why does December work better than any other month for this?
Because three things line up at once, and they only line up once a year.
First, watch behaviour changes. People are off work, on couches, on TVs, in a browsing mood rather than a searching mood. Browse and suggested traffic reward a strong thumbnail far more than search does, so a better image gets a bigger multiplier than it would in April.
Second, upload volume drops. A lot of channels in your niche go quiet from around the 18th of December. Less new supply, same demand. Your two year old video is suddenly competing against a thinner field.
Third, you finally have time. Refreshing 8 videos properly takes thought, and thought is the thing you never have in a normal publishing week.
There is a fourth reason that is less pleasant. January is when the algorithm sees a lot of new viewers with new intent, and those viewers meet your library before they meet your new uploads. If your library looks like it was packaged by a different, worse channel, that is your first impression.
Which videos actually deserve a refresh?
Start with the diagnostic, not the design. I sort the library into four buckets, and only two of them get budget.
Bucket one: high impressions, low click through. YouTube is already showing the video and viewers are declining it. This is the single best repackaging candidate in existence. The distribution is free and sitting there. If a video has thousands of impressions and a CTR well below your channel average, the packaging is the bottleneck, full stop.
Bucket two: high retention, low impressions. People who find it love it, but almost nobody finds it. Slightly harder, because you are asking the system to test it again. A new thumbnail and title gives it a legitimate second audition. Not guaranteed, but the odds are decent and the cost is one thumbnail.
Bucket three: high CTR, poor retention. Leave it alone. The packaging is working and the video is not delivering. Repackaging harder makes the mismatch worse. This is where I get honest with clients: some videos are not a design problem.
Bucket four: genuinely dated content. Software from three versions ago, a prediction that already resolved, a price that is wrong. Repackaging dated content is how you earn angry comments. Skip it or remake it.
A practical rule I use: if a video would be embarrassing to have at the top of your channel in February, do not spend money making more people see it in January.
How many videos should you refresh?
Eight is the number I keep landing on, and it is not arbitrary. Fewer than five and you cannot tell signal from noise, because any single video can move for reasons that have nothing to do with you. More than twelve and you dilute your attention, ship weaker concepts, and cannot read the results cleanly.
Eight also happens to be exactly what our thumbnail batch is built around. Two hundred and fifty dollars for 8 custom thumbnails, roughly 31 dollars each, one per video, unlimited revisions, a dedicated designer, and title consultation included. Turnaround is 48 to 72 hours per thumbnail, which means a batch of 8 fits comfortably inside a December window if you start before the holidays swallow everything.
For channels that are smaller and unsure whether this is worth it at all, the short version is this: at low view counts the library is a smaller lever, but it is also a much cheaper experiment.
What do you actually change on an old thumbnail?
Here is where most refreshes go soft. People make the old thumbnail prettier. Prettier is not a strategy.
The changes that move numbers, roughly in order of impact:
- Change the idea, not the execution. If the old thumbnail said "here is a tutorial about X", the new one needs to say something a person would stop for. A stake, a contrast, a number, a consequence.
- Cut the word count. Old thumbnails almost always have too much text. Three or four words is usually the ceiling. On a phone in a feed, six words is a grey smudge, which is the whole reason your text is unreadable on mobile more often than you think.
- Rewrite the title in the same breath. The thumbnail and title should not repeat each other. They should complete each other. If both say the same thing you have wasted half your real estate, which is the entire point of designing them together rather than treating them as separate jobs.
- Fix the crop for small sizes. Most old thumbnails were designed at full size and never checked at sidebar size. Faces too small, subject too far away, background too busy.
- Make it look current. Not trendy. Current. If your last twelve uploads share a visual language and your 2023 videos do not, the library looks incoherent to a new viewer scrolling your channel page.
If you want the official ground rules on what YouTube itself allows and rewards in thumbnails, the YouTube creator resources are worth twenty minutes before you brief anyone.
Should the refresh look seasonal?
Mostly no, and this surprises people who arrive here in December.
Seasonal styling belongs on seasonal content. If your video is a gift guide or a year in review, then yes, lean in, and the same logic I use in holiday season thumbnail strategy applies directly. Everything else should be designed to work in March, because that is where it will spend most of its life.
The exception is deliberate commercial content. If you have videos that convert, that sell something, or that feed a launch, dressing them for the season for six weeks is defensible. That is closer to packaging for Black Friday content, where the deadline is real and the payoff is measurable.
The failure mode is a channel that repackages its whole library in snowflakes and then has forty dated thumbnails on the fifth of January.
How do you know if the refresh worked?
Give it 28 days and compare against the same video's previous 28 days, not against your best video ever.
What I look for:
- Impressions click through rate, per video, not channel wide. Channel CTR moves for a dozen reasons.
- Whether impressions themselves grew. A better thumbnail earning more clicks often causes YouTube to serve more impressions, which is the real compounding effect.
- Average view duration holding steady. If CTR jumps and retention collapses, the new packaging is overpromising and you should pull it back.
Be honest about the noise floor. Out of 8 refreshed videos, a normal result is two that clearly improve, four that move a little, and two that do nothing. That is a win, because the two that improve keep earning for years and the cost was one flat fee.
If you want to see how we handle the design side across different niches and formats, our work is the fastest way to judge whether the style fits your channel. A finance channel around 80k subs and a cooking channel at 12k need very different answers to the same brief.
Is it worth paying someone, or should you do it yourself?
Do it yourself if you have taste, time, and distance from your own content. Those three together are rare, and the third one is the killer. You know why the video is good. A stranger scrolling at 11pm does not, and you cannot easily unsee what you know.
Hire it out if the maths works. Eight thumbnails at roughly 31 dollars each is cheap against the revenue of a single video that starts working again. It is also cheaper than the alternative most people choose, which is filming another video that nobody sees because it is packaged like the last one.
A few practical notes on how we run these batches. Everything is custom per video, never a template, and the styles we use get tested on our own channels first. If you want to split test a concept, an A/B variant is 15 dollars. A title pack is 25 dollars. Rush is 20 dollars, though for a December batch I would rather you start early than pay to compress it. If you want a sense of how the batch price compares to buying one at a time, I broke that down in thumbnail packs versus per thumbnail pricing.
We are based in Dubai and work async with channels worldwide, which for this particular job is an advantage. A refresh has no shoot days and no live calls to coordinate. You send the list, we send concepts, you react.
One more thing worth saying, because I see it every January. Plenty of creators bring packaging in house, get busy, and quietly let it slide until the library looks neglected. If that is you, there is no shame in it, and the honest version of that conversation is here in outsourcing again after going in house.
Where to start this week
Pull your last two years of videos, sort by impressions, and highlight every video whose CTR is below your channel average. That list is your refresh candidates. Cut it to 8 by asking one question per video: would I be happy if ten thousand new people watched this next month?
That is the whole exercise. The videos are already made. The audience already exists. The only thing left is the promise on the outside.
If you want a second pair of eyes on which 8 to pick, email me at samy@pixelsamy.com and send me the list. I will tell you honestly which ones are worth repackaging and which ones are not.