In short
Mostly no. Auto-rotating hero carousels are consistently among the least effective elements on a page: the studies that have measured them find the first slide gets the large majority of the clicks and later slides almost none, they push real content down, they cost performance, and they are difficult to make accessible. A carousel is defensible for browsing a set of equivalent items the visitor chose to look at, such as product photos or a gallery. It is not a good way to decide what your homepage says.
The short answer
If the question is about a rotating banner at the top of the homepage: no. If it is about letting someone flick through the six photos of a product they have already clicked on: that is fine, and it is a different component that happens to share a name.
The distinction is who chose to look. A visitor stepping through images of a thing they want is browsing. A visitor being shown five messages in sequence because five departments wanted the top of the page is being managed.
What the measurements say
Hero carousels have been tested repeatedly since they became fashionable, and the results have been remarkably consistent. The most-cited figure comes from a university site whose carousel drew clicks on about one percent of page views, with the first slide taking the great majority of those. Commercial tests have found the same shape: the first slide gets what engagement exists, and slides two onwards are seen by few and clicked by almost nobody.
The reasons are mundane. People act on what is in front of them when they arrive; by the time the second slide rotates in, they have either scrolled, clicked, or left. Motion at the top of the page also pattern-matches to advertising, which visitors have spent two decades learning to ignore. And anything that moves on its own steals attention from the content underneath, which is the content the visitor came for.
The costs that do not show up in click data
Performance. A carousel is several full-width images plus a script, all loaded before the page is useful, on the one page where first impressions matter most. Slide one is usually the largest contentful paint element, and the others are dead weight downloaded for nobody.
Accessibility. Auto-rotation fails people who read slowly, people using screen readers who lose their place when the content changes, and people with vestibular conditions for whom moving content is physically unpleasant. WCAG requires a way to pause, stop or hide anything that moves for more than five seconds. Most carousels do not provide it, and the ones that do rarely make it findable.
Mobile. Swipeable carousels compete with the browser's own gestures. Half the time a horizontal swipe navigates back instead. The dots that indicate position are invisible at phone size, so visitors do not know there is more.
The hidden decision. The largest cost is organisational. A carousel lets a team avoid deciding what the homepage is for. Every stakeholder gets a slide, nobody has to rank priorities, and the visitor gets a page that says five things at once, which is the same as saying nothing. The homepage headline is the single most valuable sentence on the site; rotating it means it does not exist.
When a carousel is fine
Product image galleries. The visitor chose the product and wants to see it from several angles. Manual navigation only, thumbnails visible, no auto-advance.
Logo strips and testimonials in a scrolling row. A horizontally scrolling list of equivalent, low-stakes items, where nothing important is hidden and the visitor can ignore it entirely. Static is still better on mobile.
Genuinely sequential content. Step-by-step instructions where the order matters and each step is a full screen. Rare on marketing sites.
In every case: user-initiated, never auto-rotating, with visible controls large enough to tap, and with the first item useful on its own.
What to build instead
One hero with one message. Pick the thing most visitors need and lead with it. If the answer is genuinely different for different audiences, that is a segmentation problem: two or three clearly labelled paths below a single headline, not a rotation.
A stacked layout. The slides that would have been in the carousel become sections down the page, each with its own heading and call to action, all visible to scrolling, all indexable, all measurable. This almost always performs better because every message is actually seen.
A static feature with a short list. One featured item, three or four smaller ones beside or below it. This is how every news site solved the same problem, and their layouts are tested against more traffic than yours.
If you have to ship one anyway
Sometimes the decision has been made above your pay grade. Then: no auto-rotation, or a pause on hover and focus with a visible pause control if the client insists; a maximum of three or four slides; the first slide's image preloaded and the others lazy-loaded; keyboard-navigable controls with proper labels; a prefers-reduced-motion check that disables any animation; and an analytics event on each slide's control so you can show, in a month, that nobody used it.
That last one usually ends the conversation for the next redesign. If you would rather have the conversation before building it, book a call and we will help make the case with your own data.
Common questions
Do website carousels hurt SEO?
Indirectly. Carousels load several large images and a script before the page is useful, which hurts Largest Contentful Paint and therefore Core Web Vitals. Content on later slides is often rendered in a way search engines treat as secondary. The bigger cost is that the page's main heading and message are unclear because they rotate.
Are carousels bad for accessibility?
Auto-rotating ones usually are. Moving content disorients screen-reader users, penalises slow readers, and can trigger vestibular symptoms. WCAG 2.2 requires a way to pause or stop anything that moves for more than five seconds. A manually controlled carousel with labelled, keyboard-reachable buttons and no auto-advance can meet the guidelines.
What percentage of people click on a carousel?
The most-cited measurement found about one percent of visitors clicked any slide, with the first slide taking most of that. Commercial tests have reported similar shapes: whatever engagement exists goes to slide one, and later slides are seen by few and clicked by almost none. Treat the exact figure as illustrative and test your own if you have one.
What should I use instead of a homepage slider?
One static hero with a single headline and call to action, followed by the other messages as sections down the page. Every message is then actually seen, each can be measured, and the page loads faster. If different audiences need different messages, offer labelled paths below one headline rather than rotating between them.
Is an image carousel okay on a product page?
Yes. A product gallery is a set of equivalent images the visitor has chosen to browse, which is what a carousel is good at. Keep it manual, show thumbnails so the visitor can see how many there are, make the controls large enough to tap, and never auto-advance.
