Why We Stopped Recommending Sliders

Almost every website brief we receive asks for a rotating slider on the homepage. It's understandable — it looks like a solution to a real political problem, which is that four departments each want the top of the homepage.

It just doesn't work, and it hasn't for a long time.

What actually happens

Study after study, across large sites with enough traffic to know, finds the same pattern: the first slide gets nearly all of the (already small) engagement, and slides two through five get almost nothing. Not less — almost nothing.

Meanwhile the slider is doing damage:

• It costs you speed. Five hero images, a carousel library, and often a layout that shifts as it initializes. That's your Largest Contentful Paint and your layout stability, both spent on content nobody sees.
• It moves while people read. Anyone who started reading slide one and is now looking at slide two has been actively interrupted.
• It's poor for accessibility unless carefully built — auto-advancing content is genuinely hostile to screen reader users and to anyone who reads slowly.
• It hides your message. A visitor gets a few seconds. Splitting your one clear statement across five rotating ones means they get a fraction of each.

Why it keeps getting requested

Because it resolves a disagreement without anyone losing. Every stakeholder's message goes on the homepage; nobody has to prioritize.

That's the actual function of a homepage slider, and it explains why the data never changes anyone's mind. The slider isn't solving a visitor problem — it's solving an internal one.

What to do instead

Pick one message. The single thing you'd say if you got one sentence. That's your hero. This is a hard conversation and it's the entire value of the exercise.

Put the others below it. A homepage can be long. Three clear sections stacked vertically will be read far more than five slides rotating, because the visitor controls the pace.

If content genuinely rotates, make it a grid. Multiple offers, multiple audiences, multiple case studies — show them at once. Visitors are good at scanning and bad at waiting.

If you truly need a carousel, build it properly. Manual advance only, visible controls, pause on hover and focus, keyboard accessible, real text rather than words baked into images. That's a well-built carousel, and it's still worse than not having one.

The conversation that resolves it

When a client asks for a slider, we ask a different question: if a visitor only reads one sentence, which sentence should it be?

That question is much more useful than the slider debate, and once it's answered the slider request usually withdraws itself.