The collab is one of the few moments on YouTube where you get to borrow someone else's audience, right, and it is strange how often that moment gets wasted on the packaging. Two creators spend weeks planning the video, travel to film it, edit it carefully, and then the thumbnail is two faces side by side, both smiling at the camera, with both channel names or a plus sign in the middle, and the title is something like "I Met Up With My Friend", which tells a stranger nothing about why these two people together are worth twenty minutes of their evening.
I think about collabs a lot, partly because the Pixel Samy group has channels of its own, seven of them past 100,000 subscribers each, where every style is tried on our own uploads before a client video ever sees it, so I look at packaging as someone who has to live with the results. The way we work at ClickTheory, the words and the picture are written as a pair, and with collabs that pair has three questions to settle, who leads the frame, whether each channel gets its own version, and what the title actually sells about the pairing.
Who should lead the frame in a collab thumbnail?
The person whose role in the premise is the hook should lead, and the person the audience already recognises should anchor, so the viewer sees a familiar face and a new reason to click in the same glance.
Two equal faces side by side is the default, right, and it is usually the weakest option, because the eye has nowhere to land first and the frame reads as a photo of two people rather than a moment. The stronger approach is to give one person the dominant position, bigger, closer, in sharper focus, and to let the other support the story, reacting, watching, standing behind. Which one leads depends on the premise. If the video is your guest teaching you something, the guest leads and your face reacts, and if the video is you taking your guest somewhere they have never been, you lead and their reaction carries the curiosity.
The exception is when one of you is far better known than the other, and I'm pretty sure most creators who have done a few collabs have felt this one. Then the bigger name usually leads on both channels, because recognition is the strongest reason anyone stops, and the less known creator gains far more from the association than from equal billing. The detail of where viewers actually look is covered in faces, eyes and where viewers look first, and it explains why two faces at the same size compete with each other instead of helping.
Two audiences means two thumbnails
When a collab gets posted on both channels, the same thumbnail on both is usually a mistake, right, because the two audiences are coming from different places and want different things from the video. Your audience knows you and is curious about the guest, while their audience knows them and is curious about you, so the version on your channel should introduce the guest to your viewers, and the version on their channel should introduce you to theirs.
In practice, right, that often means the same shoot and the same idea, but a different person in the dominant position, a different line of text, and sometimes a different title. It costs a little more work, and to be very honest, of everything in this post, this is the single biggest packaging improvement most collabs could make, because each version is designed for the people who will actually see it. Think of it this way, the viewer on your channel already trusts you and needs a reason to care about the guest, and the viewer on their channel already trusts them and needs a reason to care about you, so one thumbnail trying to do both jobs ends up doing neither of them very well.
What should a collab title actually sell?
The pairing itself, meaning the reason these two people together create something neither could create alone, whether that is a challenge, a disagreement, a skill swap or a question one of them can answer for the other.
"I Spent a Day With a Professional Chef" sells a pairing, "Chef vs Home Cook, Same Ingredients" sells a tension, "I Asked a Pilot What They Never Tell Passengers" sells a question only the guest can answer. Each of those tells a stranger exactly why the collab exists. What does not work is the title that only makes sense if you already know both people, "Finally Met Him", or a title that names the guest without saying what they bring, because the viewer from outside either audience has no reason to care yet.
The thumbnail then carries what the title does not, so if the title names the tension, the thumbnail shows the moment it peaks, the two faces reacting to the result, the dish, the score. That split between the title and the thumbnail is the whole idea in the title thumbnail pair, and it matters even more in a collab, because both pieces have to explain two people very quickly. On a collab I ask our strategists to draft the title list for both channels in the same pass, so the two versions grow out of one premise instead of drifting apart over a week of messages.
A collab thumbnail should answer one question for a stranger, why these two, and if it cannot, no amount of design will save it.
Where the eyes go when there are two faces
Faces are what the eye lands on first in almost any thumbnail, and with two of them the designer has to decide the order deliberately, right, or the viewer decides for you and usually gets it wrong. The face that is larger, closer to the camera and looking toward the viewer gets read first, and the second face should either look at the first one or at the object of the story, which pulls the eye along a line through the frame instead of bouncing between two portraits.
Expression matters just as much, right, because two people with the same expression make a frame static, while two different expressions, one confident and one shocked, one explaining and one unconvinced, create a small story inside the image. Our page on faces in YouTube thumbnails covers when faces help and when they hurt, and the honest answer for collabs is that they almost always help, as long as they are doing something. You see why the default two-heads layout fails so often, it gives the viewer two faces and no story.
Can you A/B test a collab thumbnail?
On a regular upload it works well, since YouTube will split viewers across up to three title and thumbnail versions and then show everybody whichever one kept people watching longest, but it does not work on Shorts, or on a premiere until that premiere has turned into a normal video.
The YouTube A/B test for titles and thumbnails runs from YouTube Studio on desktop and needs advanced features turned on, and collabs are a very good candidate for it, because the question of who leads the frame is genuinely uncertain and the test gives you a fair answer. I like to test the lead swap directly, one version with you dominant and one with the guest dominant, and let watch time decide. The results go into the channel's packaging playbook, winners and losers both, so the next collab starts from evidence rather than taste. How we run these tests for clients is on the YouTube thumbnail A/B testing service page.
How we package a collab, step by step
Collabs need a bit of planning before the shoot, because the thumbnails are only as good as the stills you capture on the day, and once the other creator has flown home you cannot reshoot a reaction.
The step most people skip is the first one, right, and the catch here is that if the two creators have not agreed what the collab sells, the stills will not support it and the titles will pull in different directions, so a short conversation before filming saves a lot of compromise later. If the collab is really an interview, the logic shifts slightly toward one guest and one claim, which is covered on our page about thumbnail design for interview shows.
What about sponsored and brand collabs?
When a collab involves a brand paying either creator, the packaging still has to be honest, and the disclosure belongs in the video and in YouTube's paid promotion setting, not hidden from the viewer.
A brand collab is still a collab, so everything above applies, but there is an extra layer. YouTube has a paid promotion disclosure setting for videos with sponsored content, and depending on where you are, advertising rules on endorsements may apply too, so treat this as general information and check the specifics with whoever handles your sponsorship contracts. From the packaging side, I would never design a thumbnail that disguises a sponsored collab as an organic one, because the audience finds out and the trust you lose costs far more than the click you gained.
What does it cost to package a collab?
Our packs hold eight thumbnails for 250 dollars, and a collab posted on both channels uses two of those eight, one for each audience, with unlimited changes and title help included, and each version ready two or three days after the stills arrive, not counting weekends.
When both of you want the video out the day after filming, the 24 hour rush, which is 20 dollars, is the extra that actually matters, and the lead swap test I described above costs 15 dollars per variant. If the other creator works with their own designer, share the premise and the title direction with them early, so both versions point the same way. The work runs from Dubai in writing, over Slack and email, which is handy when the two of you live in different time zones anyway, you can stop after any pack, and every other extra is listed on services. Our broader approach to building titles and thumbnails as one unit is on the YouTube title and thumbnail packaging page.
When you do not need us for a collab
If you do one collab a year with a friend and you both enjoy making thumbnails, sit down together with the ideas above and design your own versions, because the thinking matters more than the polish. If the other creator already has a designer they trust, let each channel package its own version and just agree the premise and the title direction, which gets you most of the benefit. And if you are looking for an editor rather than a packaging studio, that is a different job, handled on the Pixel Samy side rather than at ClickTheory.
Where we fit is the creator who collabs often enough that the packaging decides whether those borrowed audiences turn into subscribers, and who wants each collab tested and recorded rather than guessed. At the end of the day a collab is a chance to meet a whole new audience, and the thumbnail is your introduction, so it should be a good one. If you send me your last few collabs, on a 15 minute call, by email to samy@pixelsamy.com or through contact, I will show you what I would have done differently with each one and what a pack would cost you. Two audiences only meet each other for the first time once, right, so that introduction is worth doing properly. So yeah. That's my way of saying it.