When a BMX rider sends us a clip of a trick, I'm pretty sure they expect us to grab the highest frame, right, the moment the rider is furthest off the ground, because that is the frame that felt biggest when they landed it, but for most BMX tricks the highest point is actually the worst frame in the clip, because the bike is often edge on to the camera, the rider is tucked into a shape that could be anything, and at phone size it reads as a small dark blob against the sky.
The frame that sells a BMX video is the peak of the trick, which is a different thing from the peak of the air. It is the exact moment when a viewer who knows BMX can tell what trick is happening, the frame of a tailwhip where the frame of the bike is sideways to the rider, the barspin where the bars are clearly mid turn, the tabletop where the bike is flattened out, and finding it takes going through the clip frame by frame. At ClickTheory we build the image and the headline together, and for BMX the image is almost entirely about getting that one frame right, then making sure nothing else in the thumbnail fights it.
What is the peak of a BMX trick, and why is it not the highest frame?
It is the single frame where the trick is identifiable, the bike and rider make a shape that could only be that trick, and in rotations, whips and spins that moment usually comes before or after the top of the arc.
Think about what is happening in a few common tricks, right, a 360 is a rotation, so the readable frame is the one where the rider is clearly side on or facing backwards relative to the takeoff, which is rarely the top. A tailwhip is the frame spinning around the bars, and the peak is the moment the frame is fully sideways and visibly separate from the rider's legs. A barspin is all about the bars, so the frame you want is the one where the bars are crossed and the rider's hands are off them. A flair, a backflip, a can can, a no hander, each one has its own peak, and in all of them the rider's body and the bike have to form a silhouette that an experienced viewer recognises in a fraction of a second.
That is one reason I always ask BMX riders for the original clip rather than a screenshot. Secondly, high frame rate footage matters a lot, because a trick that happens across six frames at a normal frame rate has a much better chance of containing its peak if it was filmed at a higher one. And the third thing is shutter speed, because a blurry bike is the most common reason a great trick cannot carry a thumbnail at all.
Why do BMX tricks disappear against a skatepark background?
Because concrete, ramps, fences, graffiti and other riders create a busy grey background with lots of lines, and a rider in dark clothes on a dark bike blends straight into it once the image is small.
Skateparks look great in a video, right, they give a sense of place and scale, but in a still they are visual noise, and the problem gets worse the more famous and decorated the park is. The fix I come back to every time is separation, and sky is the best background a BMX trick can have, so a low camera angle pointing upward at an air over a box or a jump almost always reads better than an eye level shot with the park behind. When the sky is not available, we cut the rider and the bike out and soften, darken or simplify everything behind them, so the shape of the trick is the brightest and sharpest thing in the frame.
Fisheye footage, which is common in BMX filming, needs care, I would say more than any other lens choice. It gives a great sense of speed and closeness in the video, but it bends the ramps and the bike, and at thumbnail size a heavily curved frame can make a trick harder to read rather than more dramatic. My rule is to use it when the rider is close to the lens and the trick is clear, and we straighten or crop it when the distortion starts to fight the shape.
Do street, park, dirt and flatland need different thumbnails?
Yes, because each discipline has a different kind of peak and a different setting, and viewers in each one are looking for something specific when they scan the feed.
Park riding is about airs and transitions, right, so the sky separation approach works best, and the frame is the trick high above the coping. Street riding is about the spot, a rail, a ledge, a set of stairs, and here the setting is part of the story, so we keep enough of the obstacle visible for the viewer to understand the size of it, while still making the rider stand out. Dirt jumping is about big, smooth airs over large jumps, often in natural light outdoors, and its thumbnails can lean on the line and the scale of the jumps, though if your dirt footage comes off a helmet camera, read what I wrote for Bentonville mountain bike channels about why those frames rarely work as thumbnails. Flatland is the odd one, because the tricks happen on flat ground with the rider balancing and spinning the bike, so the peak is a balance position, the rider on one peg with the bike at a strange angle, and the frame should be shot close and low, almost like a portrait.
BMX racing is different again, it is closer to motorsport packaging than to trick content, with the gate start, the first turn and the finish as the main moments. If your channel mixes disciplines, a small consistent cue per discipline helps viewers find what they came for, which is the same system I described for triathlon channels that mix three sports in one library. Our guide for Santa Cruz skateboarding channels covers the related debate about whether the trick name belongs in the thumbnail or the title, which applies to BMX almost word for word, so I will not repeat it here.
What should a BMX tutorial thumbnail show?
The trick at a size and difficulty that matches the viewer the tutorial is aimed at, with the rider wearing a helmet, because beginners copy what they see and a learning video should never sell a big, risky version of the move.
Tutorials are some of the most searched BMX videos, how to bunny hop higher, how to manual, how to learn your first barspin, and the viewers are often young riders or adults starting late. The temptation I see constantly is to put the most impressive version of the trick on the thumbnail, a huge bunny hop over a picnic table, because it looks exciting, but that sends the wrong signal for a beginner video and it is not what the viewer can do yet. A better frame shows a clean, achievable version, low to the ground, with the body position clearly visible, and the title names the problem the learner has, "why your back wheel won't come up" rather than the trick name alone.
Safety matters here more than in edit videos, and I want to be clear this is general sense rather than any kind of advice. We keep helmets on in tutorial frames, we do not package dangerous stunts as challenges for beginners to copy, and for slam compilations we avoid frames that make a serious injury the selling point. YouTube's own creator resources are a good place to check the platform's current policies on dangerous acts if your channel films riskier content.
In BMX the trick lasts maybe half a second, right, and the whole thumbnail is about finding the one frame inside it that a rider will recognise from across the room.
How should edits, contests and bike check videos be packaged?
Edits should lead with the single best trick in them, contests with the rider and the stakes, and bike checks with the bike as the hero, because each of those videos is promising a different thing to a different kind of BMX viewer.
A full park or street edit has many tricks, right, but the thumbnail can show only one, so pick the most recognisable and impressive peak, not the last trick of the edit and not a collage of three. Contest videos and event recaps are about people, so the rider's face and reaction after a run, or the moment of the big trick on the course, carry more than any wide shot of the crowd. Bike check and build videos are the one place the bike can be the subject on its own, shot clean against a plain background with the parts that are new or unusual visible, and the rider small or absent. Collab videos with a group of riders work best with one rider's trick and a small row of faces, rather than four tiny riders in the air.
Brand and shop channels, the ones selling frames, parts and complete bikes, can use the same thinking, the trick frame for their team videos and the clean product frame for their product content, and our page on sports channel thumbnails goes wider on how sports brands package. Riders who also film parkour or other urban sports may find our parkour and freerunning thumbnail guide useful, because the problem of a body in the air against a busy city background is almost the same one.
How much does BMX packaging cost, and when is it not worth it?
I charge BMX channels the same as everyone else, 250 dollars covers 8 custom frames, so a park edit or a tutorial costs you a little over 31 dollars to package, and that price already has unlimited changes and a second pair of eyes on your title in it, with frames usually landing in your inbox within three working days, weekends excepted.
For a BMX channel that posts a park edit every week or two, a pack lasts a couple of months, and because there is no retainer, a winter break or an injury layoff costs you nothing, right, you just pick it back up when you are riding again. If an edit is ready and you want it up fast, the rush option gets a thumbnail back inside a day for 20 dollars extra, the other add-ons are on services, and I explain who actually designs your frames on the page about hiring a designer from a studio that runs its own channels. For riders whose channels are growing, the A/B test variant is worth using on edits, because two different peak frames from the same video can be tested against each other through YouTube's own title and thumbnail test.
When do I think it is not worth it? If you are filming for fun with friends and posting a few edits a year, learn to step through your own footage and pick the peak, it is a skill that costs nothing and makes a big difference. If your footage is blurry or filmed from far away, spend the money on a better camera setup first, because no designer can fix a trick that was not captured clearly. Quick check before you go, have you ever actually stepped through one of your best clips frame by frame to see where the trick becomes readable?
The trick already happened, the frame is a choice
Every BMX rider has a clip of a trick that took weeks to land and a thumbnail that makes it look like nothing, and the difference is almost never the riding, it is which frame got chosen and what was left around it. Choose the peak, not the top, separate the rider and the bike from the park, keep the frame honest about how hard the trick is, and the work you put into landing it finally shows up in the feed. So yeah. That's my way of saying it.