On the homepage above the fold, avoid carousels altogether.

A quickly rotating image distracts attention away from the CTA button.

What I usually recommend is to either have a static background image or a slowly moving background video, especially if it is well-integrated with the text & CTA(s) above.

I wrote an article on my blog about the Pros and Cons of a Homepage Background Video.

By a ‘slowly moving background video’, I’m thinking of something like this, for example:

The moving water looks cool, attracts your attention for a second, then you actually focus on the headline above it.

 

Opt instead for user-initiated carousels, if any.

If you are to have an automatically or manually rotating hero image carousel, this is one way to make it work:

Keep the elements on top of the image fixed (headline, subheadline, CTA) and only change the background image.

 

My research on this

1. A case study by Grizzlyzoos showed that a homepage with a carousel received 2.06% clicks, while having a static image received 40.53% clicks. (source)

2. In the experiment on ND.edu‘s carousel, 84% were on stories in position 1 with the rest split fairly evenly between the other four (~4% each). So the other slides besides the first one are basically ignored.

 

 

 

 

3. More experiments were done on Notre Dame sites, with these results. Translation, people click a lot more on the first slide of the carousel than subsequent slides.

 

 

4. Auto-rotating carousels annoy users and reduce visibility (source)

 

5. Auto-rotating carousels get ignored. (source)

 

 

6. Carousels are missed by users (source)

 

 

7. A funny illustration of why carousels suck: http://shouldiuseacarousel.com/

 

8. In VWO’s words… ‘Don’t use image sliders’. (source)

 

9. Chris Goward’s testing of rotating carousels. (source)

 

 

10. Nobody clicks on carousels. (source)

 

 

11. A static image converts much higher (source)

 

 

12. Mostly the first slide is clicked. (source)

 

 

13. A/B tests by ServerTastic showed that versions without sliders outperformed those with sliders by an average of 16.48% in terms of revenue per visitor. (source – the original study page isn’t available anymore)

 

14. By optimizing the order of the slides in the slider, Conversion Sciences was able to put together a page with a slider that outperformed any static image by 61%. (source)

 

Examples