Most of us are familiar with the carousel, even if we don't know it by name. You may know it as a slideshow or a marquee, but they all have the same functionality. It's generally the first thing a user sees when visiting a site - a big block of rotating content that lies at the top of the homepage. The goal is to show users as much content as possible, while giving them a call to action.
These are just a few examples of carousels that iMarc has built for clients. We often include them in sites as they are a common request. But just how effective are these carousels at providing meaningful content and giving a call to action that the users actually take?
Why do we use carousels?
The use of carousels is derived from a need to give multiple pieces of content a center stage on the homepage. Many clients believe it is important to have the most important content lie "above the fold." Generally, this is driven by a need to satisfy multiple departments within a company wanting their content front and center.
From a developer's point of view they are easy to implement due to the abundance of ready-built plug-ins. The combination of ease-of-implementation and the demand from clients makes carousels a very popular tool in the web development industry.
What does the data tell us about carousels?
Unfortunately, most carousels have proven to be less than effective at engaging users. Erik Runyon has some great data reflecting this conclusion.
- User engagement with the carousel ranged from 1.07%–9.409%. The high end of the range was an outlier, with the majority of the carousel engagement being in 1–2% range.
- Within that 1–9% range, the first slide in the carousel was clicked 54.57%–89.1%. Subsequent slides had engagement as low as 2.4%. This mean that slides were engaged by as low as .02568% of users. You can see why this could be a problem.
- On the bright side, it seems randomizing the order of the slides helps with engagement. If the first slide is different everytime a user lands on the page, then the engagement should be spread out equally across all slides.
- Using a carousel that automatically switches between slides raises engagement. The site that used an autoplaying carousel had the highest engagement, at 9.409%. It also had the highest engagement for subsequent slides, which ranged from 7.97%–17.87%. Both of these figures are significant deviations from the other sites tested.
What about iMarc specific data?
Recently, iMarc conducted user testing for the a client's website. After providing the users with a series of tasks to complete, we asked them if they could recall what the three slides on the homepage carousel were about. Zero of the testers were able to list back all three correctly, and the majority were only able to recall the first slide. Further, we found the less content that was on a carousel slide, the better. This is especially important if the carousel auto-scrolls – users must be able to quickly scan the content before the slide transition occurs.
What should we take away from this data?
Before I wrote this post, I was ready to throw out the idea of carousels all together. Having looked at the data, I still believe they are not very effective at engaging users. However, if you must use a carousel there are several paths to improvement.
- Randomize the order of the slides for every user when they land on the page. This prevents the first slide from being the only content the user engages with.
- Turn on autoscrolling. Many users have difficulty finding or using the navigation controls for carousels. This prevents them from ever seeing anything but the first slide.
- Continually refresh the content in your carousel. Often times the carousel is used in a prime spot for promoting new material. If content is left to linger, it can become stale, wasting real estate that could be used more effectively.
Lastly, I'd like to point to Dave Tuft's post on "above the fold" which is a good complement to this post. The use of carousels and the emphasis on above the fold content for webpages are two common points of contention. It may be time to rethink both.