Watch fifty pieces of online video from different creators in different subjects and a surprising amount of it looks like the same thing. Similar colour, similar framing, similar rhythm, similar sound.
Imitation explains part of it. The larger part is tooling, and that is a more interesting story.
Presets are a shared aesthetic
Most people do not grade from scratch. They apply a look — a preset, a lookup table, a filter — and adjust.
There are a limited number of popular ones and they are used by an enormous number of people. Which means a very large amount of video is passing through the same handful of colour transformations.
The teal-and-orange separation that dominated for years is the obvious example, and the current fashion for lifted blacks and desaturated midtones is the same phenomenon with different parameters.
None of this is a criticism of the presets, which are well made. It is an observation about what happens when a good tool becomes ubiquitous.
Default settings are enormously influential
Every piece of software ships with defaults and most people do not change them.
Default transition durations. Default text styles and animations. Default caption timing and positioning. Default sharpening and noise reduction. Default loudness targets.
Each of those is a small aesthetic decision made by a developer and reproduced across millions of pieces of work.
The caption style used by one popular editing application is now so widespread that it reads as a format convention rather than as a software default, which is what happens when a default persists long enough.
Music libraries are small
Licensed music for online video comes overwhelmingly from a handful of subscription libraries.
Each has a catalogue of a few thousand tracks, of which a few hundred are surfaced by the search and recommendation within the library, of which perhaps fifty are used constantly.
Which means the sonic character of an enormous amount of online video is determined by what a small number of library curators chose to commission and promote.
You can hear this. Certain drum sounds, certain chord progressions, a particular kind of plucked string ostinato, all of which were fashionable within library music at a particular moment.
Stock footage does the same thing visually
The same argument applies to imagery. Stock libraries have finite catalogues, search algorithms that surface a subset, and licensing that permits unlimited reuse.
Which is why certain shots appear everywhere. The same aerial of the same coastline. The same conceptual shot of hands on a keyboard.
For anything using stock heavily, the visual identity is being assembled from a shared pool, and the pool is smaller than it appears.
Performance feedback pushes toward the middle
The mechanism that ties it together.
Everyone can see what performs. Things that perform get imitated. Imitations that perform get imitated further. Anything unusual performs unpredictably and gets abandoned faster.
Which is a selection process with a strong pull toward whatever is currently working, and the result is convergence.
The pacing convergence is the clearest case. Cutting rates across online video have risen steadily and consistently, and there is no aesthetic argument behind it. It is a response to a retention metric, applied by everybody at once.
Where the divergence still is
Not everything has converged, and where it has not is informative.
Anything shot in a distinctive real place looks distinctive, because locations cannot be standardised.
Anything driven by a specific person's manner is unmistakable, because presence does not come from a preset.
And anything with an unusual structure stands out immediately, because structure is the layer nobody has automated.
Which suggests where to spend effort if the sameness bothers you. Not on finding a different colour preset, which is a different position within the same shared space, but on the parts that no tool has an opinion about.
Whether it matters
An honest question. Standardisation raises the floor considerably. Video made by beginners today looks vastly better than equivalent work fifteen years ago, and the tools are why.
What is lost is the top of the distribution, or rather the distinctiveness of it. A competent standardised look is easier to reach and harder to exceed, because exceeding it means abandoning tools that are doing real work.
My own compromise is to use the tools for everything technical and to make the structural and locational decisions myself, on the grounds that those are where difference is still cheap.
The archive comparison
A test worth doing if you want to see the effect clearly. Find video made by amateurs fifteen or twenty years ago and watch a stretch of it.
It is worse in every technical respect. Exposure is wrong, audio is poor, the colour is whatever the camera decided, and the editing is loose.
It is also wildly various. No two pieces look alike, because there were no shared tools imposing a look and everybody was solving the problems badly in their own way.
Whether you find that charming or simply bad probably depends on whether you had to watch it at the time. What it demonstrates is that the sameness is a recent condition and a consequence of good tools rather than of bad taste.