Web animation gets argued about as though it is one thing you either use or avoid. It isn't. Some motion earns its place on a page and some actively costs you visitors, and the difference is usually whether the animation is doing a job or just existing.

Here is how I decide.

Animation that does a job

Motion works when it answers a question the visitor already has.

A button that responds when you hover confirms it is clickable. A form field that shifts when you select it tells you where you are. A progress indicator that moves tells you the thing is working and you have not broken it. None of these are decorative. They are answers.

The same applies to explanation. If a product does something hard to describe in a sentence, showing it in motion is faster than any paragraph. That is most of why explainer videos exist.

Animation that gets in the way

Motion costs you when it delays what someone came for.

Scroll-triggered fades are the usual offender. One or two feel considered. A page where every element waits its turn to appear means a visitor who scrolls quickly sees a screen of nothing, twice. They came to read something and the site is making them wait for permission.

Autoplaying background video is the other one. It looks impressive on a fast connection on a desktop, and it is a wall on a phone with three bars.

The performance question

Every animation has a weight, and weight becomes load time. Load time affects both whether a visitor stays and how the page ranks.

The practical version: prefer CSS transforms over animating layout properties, keep video compressed harder than looks necessary, and test on a real phone on mobile data rather than on your own machine on wifi. Your machine is not representative and never has been.

Accessibility is not optional here

Some people get motion sickness from parallax and large scroll effects. The browser already has a signal for this, prefers-reduced-motion, and respecting it takes very little work. When someone has that setting on, cut the large movements and keep the small state changes.

This is worth doing regardless of how many visitors you think it affects, because the cost of supporting it is close to zero and the cost of ignoring it is that a portion of your audience physically cannot use the site comfortably.

A test that works

Turn every animation off and look at the page. If it still makes sense, still guides you to the right place, and still feels like the brand, then the animation was doing a job and you can put it back.

If turning it off makes the page better, that is your answer.