A puzzle video is a strange thing to sell in one still frame, right, because the whole pleasure of it is watching something messy become solved, and a still can only show one side of that. The cube is scrambled or it is finished, the puzzle box is closed or it is open, the jigsaw is a heap of pieces or a picture, and the viewer scrolling past has to feel the gap between the two without seeing it happen, and the most exciting part of a speedsolve is a blur of fingers that lasts a few seconds, which is close to useless as a photograph.
What makes puzzle channels easier than they look, though, is that the stakes almost always come down to one of two things, the challenge (can this be done at all) or the time (how fast, or how long it took), and once the packaging commits to one of those the frame mostly designs itself. That is the lens we use at ClickTheory for cubing, puzzle box, jigsaw and logic puzzle channels, and the rest of this is how it plays out in the feed.
What does a puzzle thumbnail actually have to show?
It has to show the puzzle in its unsolved state plus one clear signal of the stakes, a time, a size or a difficulty, so the viewer understands the challenge before they understand anything else.
Cubing channels love a frame of the creator grinning over a finished cube, right, and that frame tells the viewer that something was solved and gives them no reason to watch it being solved. Compare that with a scrambled 7x7 held up next to a face that looks slightly worried, or a sealed wooden puzzle box with "day 3" written beside it, and you can feel the difference straight away, because the second kind of frame leaves a question open that only the video can close. The challenge is the product, so the frame has to put the viewer at the start of it.
I'm pretty sure this is why cubing tutorials and cubing attempts need different packaging, too. A tutorial viewer wants to reach the solved state themselves, so the finished cube is a promise to them ("you will get here"), while an attempt viewer wants to watch someone else struggle, so the scrambled cube and the clock are the promise. We look at which of those two viewers a video is for before anyone opens a design file.
How big should the solve time be on a cube thumbnail?
If the video is about speed, the time should be the largest text in the frame and set in clear, even digits, because a solve time is the one piece of text a cubing viewer reads before anything else.
Cubers read times the way runners read splits, right, and a number like 9.84 or "sub 15" carries a precise meaning to anyone who solves, so it can do the job of a whole headline. That means it gets real space, thick numerals that do not blur at phone size, and nothing else in the frame competing at the same size. General type sizing for phones is something I covered in the post on text that falls apart on mobile, and for timer digits in particular, the main thing is to avoid thin fonts where a 1 and a 7 start to look alike.
There is an honesty rule here that I hold firmly, right, the time on the thumbnail has to be a time that actually happens in the video, and it should not be called a record unless it really is one by whatever standard the video claims. Cubing viewers know the difference between a home solve and an official result, they know what an average means against a single, and they will point it out in the comments within minutes. A channel that inflates a time once loses more trust than the click was worth, so if the best time in the video is a single and the average is slower, the thumbnail says which one it is showing, or the title does.
Why do cube colours fight with thumbnail text?
Because a standard cube already carries six loud colours, white, yellow, red, orange, blue and green, so any text or background in one of those colours blends into the stickers and disappears at feed size.
This is a problem almost no other niche has in quite this form. A designer's instinct is to put a yellow headline on a dark background, and then the yellow face of the cube sits directly beside it and the two merge into one bright patch, or a red arrow points at a cube with a red face toward the camera and simply vanishes. What we do on cube thumbnails is basically treat the cube as the colourful object and keep everything else quieter, a near neutral background, text in a colour that is not one of the six (or white with a heavy dark outline), and the cube turned so its busiest faces are toward the lens. The colour of the creator's shirt matters more than people think too, since a bright green hoodie behind a cube with a green face is one more patch of noise.
Scrambled or solved, which cube belongs in the frame?
Scrambled in almost every case for attempts and challenges, solved for tutorials and reviews, and the occasional split of both only when the video's whole idea is the transformation itself.
A scrambled cube in focus reads as a little block of chaos, which is exactly the feeling an attempt video wants, but a scrambled cube out of focus reads as a smudge, so the still has to be shot on purpose, sharp, close and well lit, rather than pulled from a frame of the solve. For speedsolves I ask creators to shoot a few stills during inspection, before the timer starts, with the cube held toward the camera and the timer visible at zero, and then a few more at the instant they stop the clock, hands open and the solved cube on the mat. Those two moments are the only ones in a fast solve where the hands are not a blur, and they give us the before and the after to choose from.
For big cubes and odd shapes, right, the 6x6 and 7x7, the mirror cubes and the custom mods, the shape of the puzzle is part of the stakes, so it needs to be large in the frame, and a hand holding it gives the viewer the scale they need to understand that this is not a normal cube.
Puzzle boxes, jigsaws and the spoiler problem
Mechanical puzzle and puzzle box channels have the opposite problem to cubing, because the solution is the secret, and a thumbnail that shows the open compartment or the hidden mechanism has given away the reason to watch. So we keep the puzzle closed in the frame, we show the creator's hands on it or their face looking at it, and the stakes go in a small piece of text, the time spent, the number of steps, "nobody has opened this" if that is true for them. The open box is for the end of the video and for nothing else.
Jigsaw channels sit somewhere in between, right, and the table full of pieces is the hardest frame in the whole niche, because a thousand small pieces at feed size is texture and nothing more. What works better is a tight crop of the half finished puzzle with an obvious gap, or the finished picture shot straight down with the time written over a quiet corner. The printed image on the puzzle belongs to whoever made it, so I prefer frames where the creator, the gap and the clock lead and the artwork is just context.
Logic puzzles on a grid, sudoku and the like, run into the same readability wall that chessboards do, and rather than repeat it here I would point you to how we handle a full chessboard at feed size, since the trick of zooming in on the one critical area works almost exactly the same way.
A finished puzzle answers the question, a scrambled one asks it, and the thumbnail is only ever there to ask.
Tutorials are packaged for search, attempts for the feed
A beginner cube tutorial lives on search, right, people type "how to solve a Rubik's cube" or the name of a method, so the title should carry those words plainly and the thumbnail should look calm and achievable, one clean cube, a friendly face, and maybe a short phrase about ease. Attempt and challenge videos live on the home feed and in suggested videos, next to other cubers, so they need the tension and the number, and their titles can afford to be a question or a goal ("going for sub 10").
Record attempts and personal best chases share a lot with timed gaming, so if your channel sits between cubing and speedrunning, our notes on speedrunning thumbnails cover the timer culture from the gaming side, and the general rules for selling an attempt that might fail are in packaging challenge and experiment videos. Channels that teach the maths behind the cube, group theory and the like, will find their closest neighbour in our math channel thumbnails.
Shorts and the first frame of a solve
Short solves are made for vertical video, right, a scramble, a burst of hands and a solved cube in fifteen seconds, and for Shorts the first frame matters about as much as a thumbnail does on long form, because it is what the viewer sees before deciding to stay. We set the first frame as the scrambled cube held to the camera with the timer at zero visible, so the clip opens on the question rather than the middle of the blur. Our page on Shorts cover design has the vertical layout rules, and for cube clips the main adjustment is keeping the cube in the upper middle, clear of the captions and buttons along the bottom.
When does a puzzle channel need ClickTheory, and when does it not?
When you post often enough that a consistent look matters and you would rather spend your hours solving than designing, and not when you post a few videos a year and your stills are already strong.
Can I ask you to do one thing before deciding? Line up your last eight thumbnails and check how many show a solved cube or an open box, because if most of them do, the cheapest fix is to reshoot your stills at inspection time and you may not need anyone. If you want the work done for you, our sets are 250 dollars for 8 thumbnails with title help and revisions included, so a little over 31 dollars a video, back in two to three working days, and the details are on our services and add-ons. For tutorial libraries that have aged badly, the title and thumbnail test inside YouTube Studio is a good way to try a calmer frame against the old one before you commit.
A puzzle channel always has the same story, something unsolved meets someone stubborn, so every thumbnail only needs to show the unsolved thing and the stubborn person at the start, with the clock still running.
So yeah. That's my way of saying it.