A museum or a library has the one thing almost every YouTube channel would kill for, which is a store of real, strange, beautiful objects that nobody else can film, right, a medieval manuscript with a doodle in the margin, a fossil with a bite mark, a letter written the night before a battle, and then the video about that object goes out with a thumbnail of a quiet gallery, the exhibition title in a thin serif and the institution's logo in the corner, and it gets watched by the people who already follow the channel and almost nobody else. I see this pattern so often that I think it is the default, and the reason, to be very honest, is that institutions package their videos the way they design their wall labels, respectfully, accurately and at a distance.
That respect is not the problem. The problem is that a wall label is read by someone already standing in front of the object, while a thumbnail has to convince someone scrolling past to care in the first place. At ClickTheory, the title and thumbnail team within Pixel Samy, our whole argument for cultural institutions is that curiosity is the right register, not solemnity, and that you can be curious without being cheap.
Why do museum and library videos look so solemn in the feed?
Because they inherit the visual language of the building, quiet galleries, formal portraits, exhibition branding, which signals importance but gives a scrolling viewer no question to answer and no reason to click.
There are a few habits behind it. Exhibition videos often reuse the exhibition's poster or key art, which was designed for a banner on a street and not for a small rectangle on a phone, so the type turns to dust and the image loses its point. Talks and lectures get a photo of the speaker at a lectern, which looks like every other lecture ever recorded. And collection videos lean on a wide shot of a room or a display case, which is how a visitor sees the object but not how a camera should sell it.
Underneath all of that is a reasonable worry, right, that anything punchier will look like clickbait and damage the institution's standing. I understand that worry, and the answer is not to copy the shocked faces of entertainment channels, it is to find the genuine curiosity in the object and put that on the frame.
What does curiosity look like on a museum thumbnail?
It looks like one specific, surprising detail of a real object, cropped large, with a short honest question or phrase that makes the viewer want to know the story behind it.
Curiosity comes from specificity, right, so "Highlights of the medieval collection" is solemn and vague, whereas a tight crop of a cat's paw print across a page of a manuscript, with the words "who let the cat in?", is specific, true and almost impossible not to wonder about. The detail does the work, and the text only points at it. For a history object, the frame might show the single line of a letter that changes how you read the whole thing. For natural history, the tooth, the claw, the scar. For art, the tiny figure in the corner of a big painting that most visitors walk straight past.
The human element helps too. A curator or librarian holding the object, looking at it with genuine interest rather than posing for the camera, reads as an expert sharing something they love, which is a very different promise from an institution presenting itself. When and how to use a face is covered on our page about faces in YouTube thumbnails, and for cultural content the rule I follow is that the object stays the star and the person is the guide.
Which object should go on the frame?
The object with the best story you can tell in one glance, which is often not the most valuable or famous piece in the video, but the one with the most surprising detail.
This is where the strategist earns their keep, right, because the institution's instinct is usually to lead with the most important item, the star loan, the most valuable book, the piece on the poster. Sometimes that is right, especially when the object is famous enough that people are searching for it by name. But for most collection videos, the better choice is the object a viewer did not know existed, or the detail of a famous object they never noticed. History channels face the same choice every day, and our page on thumbnail design for history channels covers how they pick the one image that makes a past event feel urgent, which translates directly to collections. For family history and archive material, the post on genealogy and family history channels is a close companion.
Who owns the image of an object in your collection?
It depends on the object, the photograph and the agreements around them, so a thumbnail should only use images your institution has cleared for promotional use online, and that check belongs with your rights and reproductions team.
What follows is general information from experience rather than legal advice, and your institution's own rights staff and counsel come first. The catch here is that owning an object is not always the same as having the right to use every image of it. A work by a living artist, or one still in copyright, may need the rights holder's permission for promotional use. A photograph of an object can carry rights of its own, depending on who took it and where. Objects on loan often come with agreements that limit how they can be pictured and promoted. Archives can hold material with donor restrictions. In a library, a book's cover art usually belongs to the publisher or the illustrator, not the library. And some communities have asked that certain cultural or sacred items not be shown at all, which is a matter of respect before it is a matter of rules.
So our working rule is simple, we only design with images the institution has confirmed are cleared for this use, and when in doubt, we use an object that is clearly in the public domain or one the institution owns outright, along with the institution's own photography. If your organisation is a charity as well as a cultural body, the attention versus dignity question is covered in our post on thumbnail design for nonprofits.
The object already has a story worth clicking for, so the frame only has to stop hiding it.
How do you turn a collection scan into a thumbnail?
You make a separate, prepared export from the master scan, cropped to the detail, sized for YouTube and saved as a JPG or PNG, because a museum master file is far too large and in the wrong format to upload as it is.
Collections teams produce wonderful high resolution scans, often as huge TIFF files, right, and those are not what YouTube accepts for a custom thumbnail. YouTube asks for JPG or PNG, recommends 3840 by 2160 pixels for a standard video, needs at least 640 pixels of width, and caps the file size depending on how you upload it. So the thumbnail is always its own small export, built from the scan but cropped, colour checked for a phone screen and tuned so the detail reads at feed size. The full specs are on YouTube's custom thumbnail help page.
Accessibility deserves a mention here too, because cultural institutions usually take it seriously on their websites. A YouTube thumbnail has no alt text field, so the title has to describe what the video is about on its own, and the object images you publish on your own site alongside the video should follow the W3C's guidance on accessible images. On the frame itself, large readable text and strong contrast matter far more than delicate typography.
How should libraries package talks, programmes and how-to videos?
By leading with the idea or the skill rather than the event format, so an author talk is packaged around the most interesting thing the author said, and a how-to session around what the viewer will be able to do.
Libraries publish a wonderfully mixed set of videos, right, author conversations, local history talks, digital skills workshops, storytimes, tours of special collections, and each wants a different promise. An author talk packaged as "an evening with" plus a name only works if the name is famous; otherwise the frame should carry the most surprising idea from the conversation. A digital skills workshop should look like a tutorial, a screen, a task and the outcome, because people search for the task. Storytimes are aimed at young children and families, which may mean they are made for kids under YouTube's rules, a decision that belongs to the library and changes what features the video can use. And special collections tours are museum videos in all but name, so everything above applies to them. Libraries attached to a campus can borrow from our post on thumbnail design for universities, which deals with the same mix of lectures and public programmes.
What should an institution do with its back catalogue of lectures?
Repackage the best of it, because years of recorded talks and collection videos are still being searched, and a better title and frame on a strong old video costs a fraction of producing a new one.
Most institutions I talk to have a library of hundreds of recorded talks with almost identical thumbnails, the speaker at the lectern, the series name, the date. Some of those talks hold genuinely fascinating material that nobody finds because nothing on the frame suggests it. Going back through them, picking the ones with real substance, and giving each a curious, specific frame is often the most efficient packaging work an institution can do, and it is exactly the job described on our page about redesigning old YouTube thumbnails. You see the opportunity here, right, the hard part of the work, the research and the recording, is already done.
Does a museum or library need a studio like ClickTheory?
Not always, since an institution posting a few videos a year can do well with a careful in house template, but once video is a regular part of your public programme, a studio that works on the curiosity of each object is worth what it costs.
For an institution with a regular channel, a pack of eight custom frames costs 250 dollars, each one built around its own object and story, with revisions until your curators and comms team are both happy, and usually ready within three business days, weekends excluded. If you are repackaging old lectures, testing a new frame against the old one in YouTube's A/B tool is worth the 15 dollar variant on the videos that still get impressions. The rest is on our services page. Where we are the wrong choice, I would say so: if you need on site photography of objects, a local photographer working with your collections team is essential and we do not do that, and if your rights situation is unclear for most of the collection, sort that out first, because no design can fix an image you are not allowed to use.
The objects were never the problem, and neither is the care institutions take with them. The frame just needs to share the curiosity that made someone collect the object in the first place. So yeah. That's my way of saying it.