Kill Conversion Killing Carousels Now


Kill your conversion killing carousel now, before further damage is done to your website ROI and revenue

Kill conversion killing website carousels nowKill conversion killing carousels now, because a carousel is one of the major reasons why your conversion is much weaker than it could be.

Website carousels or sliders were all the rage a few years ago. You almost couldn’t visit three sites without finding two that were using them.

What’s a carousel? It is the technology that allows a series of images to briefly appear on the home page, then rotate in order, being replaced by the next, and the next, after several seconds go by.

The theory behind using carousels was threefold:

  1. Allow multiple messages designed for multiple personas to appear on the all-important above the fold screen real estate on the home page
  2. Provide a mix of messaging in one place, typically including Branding, Product-specific, and Thought-leadership
  3. Placate internal stakeholders who demand their messaging be present on the home page

The Data is in, Carousels are Bad for Conversion

Sadly, carousels just don’t work. Not at all. Based on website audits, conversion data and usability testing I have been collecting over the past several years I can conclusively say most carousels are hurting conversion, some modestly, and many severely.

The reason carousels do not work is because the theory behind carousels is wrong. The theory is that home page visitors will hang around long enough to see each of the messages. In fact, the vast majority of website visitors will only spend a few precious seconds on a home page before either navigating into the site, or leaving it. They typically never see all the carousel images.

If a carousel has 5 images, each of which appears for 3 seconds, and allowing 1 second for the ‘sliding in’ and ‘sliding out’ transition effects, then for a visitor to see all 5 sliders it would require a total of 20 seconds (5 images X 4 seconds per image = 20 seconds total).

The problem is, website visitors do not actually stay nearly that long. Most sites are lucky if the majority of their visitors stay longer than 10 seconds.

But that is not the worst part, which is that even if they stay, often they will be more confused, not less, by the multiple messages displayed in the series of carousel images.

Going even further, many of the usability tests I conducted revealed that ‘banner blindness’ was occurring on the carousel itself. Meaning most of the study participants simply ignored the sliding or animating messages as they hunted for the information they were interested in. The proof is in the Click Through Rate of banner images, which is not good typically.

Carousel Click Through Rate below .1%

Click here to view latest job postings

Among the hundreds of website audits I have conducted in the past several years, I have seen average Click Through Rates (CTRs) of less than .1% across thousands of carousel banner images. In fact, that rate is just as bad as the average CTR for banner ads as reported by Google, which currently is at .089%.

DoubleClick click through rates for display ads image from
Google DoubleClick Display Benchmarking Report, U.S., All Verticals, All Formats, Feb 2014-Oct 2010

 Length of Visit Data is the Nail in the Coffin for Carousels

The length of visit data from the hundreds of sites I’ve analyzed over the past several years is the nail in the coffin for carousels. The data is clear and damning to carousel believers that maintain visitors will hang around long enough to view each image. In fact, as the data clearly shows, they don’t.

The average length of visit for most sites is typically under 10 seconds for the vast majority of visitors. This means that most visitors are in fact not hanging around to watch each of the carousel images advance across the home page, and are either abandoning the site immediately or moving on without the opportunity of ever being exposed to the messages in the carousel.

The image below demonstrates an average length of visit report, in this case for a client of mine. Almost all length of visit reports I have seen replicate this data fairly closely. Note that the majority of site visits are less than 10 seconds.

Visit Duration of a typical website image from
The vast majority of length of visits for most websites is well under 10 seconds

Carousels Often Fail the 5 Second Test

As I wrote about in my article on 5 second tests, a home page must communicate three critical pieces of information in 5 seconds, else losing the website visitor potentially forever. They are…

  1. Who you are
  2. What product or service you provide
  3. Why your visitor should care, how can you help them?

Carousel images are notoriously bad at passing the 5 second test. This is because although some images may be on topic, and may help the visitor understand who you are, what you do and why they should care, many other carousel images don’t. The images that fail typically counter-act any benefit derived by the images that succeed in communicating with the target audience. And often, none of the images pass the 5 second test.

Below are results of a 5 second test of two carousel images from an Ecommerce website that sells watches. One image was fairly on-point, and thus scored fairly well. The other image caused great confusion to the visitors, and thus completely failed the 5 second test.

5 second test results from a typical website carousel image from
A 5 second test can reveal which images in a carousel are not working and thus causing poor conversion

The results of the above 5 second test are very typical for the vast majority of sites I have tested. Typically out of five images, only one or two actually do the job of communicating fairly effectively. And the other three or four images are so bad at communicating that they more than counter any slight benefit gained from the good carousel images.

Conversion Improves When Carousels are Killed

Among the hundreds of website audits I have completed in which carousels were causing poor conversion, when my clients killed their carousel, they typically increased their conversion significantly.

The message is clear, kill you carousel before it kills your website!

Conclusion: Kill Conversion Killing Carousels Now!

Kill your conversion killing carousels now, before more damage is done to your website ROI and revenue. The good news is among over 50 websites I sampled, slightly less than a quarter are still using carousels, with the vast majority either removing them entirely, or using a modified version where the carousel is below the main home page image and message.

By removing your carousel and replacing it with content that passes the 5 second test, you will be better off converting those all important website visitors, which will improve your website ROI and revenue.


  1. OK, this is an interesting post, but from a UX and consulting standpoint, it misses a number of key points, and by focusing on the UI and “click-throughs,” it risks using data like a club to draw conclusions that the data doesn’t indicate.

    The first point that reads off to me is the assumption that ANYONE thinks any carousel items beyond the first will actually be seen, or “should” generate conversions. All good UXers know about the 10 second bounce rate, and don’t have expectations otherwise. This generally happens with all pages, esp. if they’ve put a lot of stock in their “hero” doing a “heroic” job. Frankly, the hero is rarely up to that task, whether it is a carousel or not. A hero should drive users to engage and take links on the site, but the links do not have to be within the carousel just because it holds the biggest chunk of real estate.

    A carousel is just a cheap way of fudging A/B/C/D testing, to tell you the truth, and a poor method of doing that as well, since data will always be overweighted on the first slide. A HERO gives space to brand and message engagement, but does not necessarily HAVE to be the item that drives click-throughs. Brand awareness, and giving a site “placeness” is far more valuable than using a big chunk of real estate just to drive click-throughs, like a massive billboard, smack-the-monkey, what-have-you.

    And this data doesn’t reveal what alternative data would be, given most web realities. Meaning, a carousel is a stakeholder negotiation tool. Key messages that would have been discarded out of hand by bad business decisions can be inserted into a carousel to, at the very least, get the message out somewhere.

    Because if you turn the carousel into a single message/image hero, I can tell you exactly what will happen. The “placate the stakeholders” message will be the only message shown (I should know. I live with those stakeholder decisions daily). And it WILL NOT drive conversions, I guarantee you. But that comparative data won’t show up in any studies.

    A carousel is a political negotiation on a site, in all instances (unless it is dynamically populated with informative and valuable news feeds, which ought to get more space in all instances, imho). A carousel helps designers hedge clients against their own bad web decisions. By rights, alt versions of a 3-slide carousel should be tested, A/B/C, with a different one of the three in the front-loaded “hot spot”.

    But to damn carousels with nothing but conversion data is like blaming the carousel UI for the poorly-chosen stakeholder branding and conversion strategies that often live inside the carousels.

  2. Thank you for your insightful article and for sharing your data! I find that carousels often serve as eye candy for designers & execs bored of looking at their own websites. The fact that they are counterproductive for users needs to be more insistently stressed. This article will help the argument! THANK YOU!

    ILONA POSNER, User Experience Consultant & Educator

  3. Thanks for your comments Chris,

    I think you raise several interesting points.

    However, I would have to disagree with you on anyone thinking carousel items beyond the first will actually not be seen. In fact, about 80-90 percent of my clients not only expect their visitors to see them, they expect their visitors to engage with each of them! I know, astonishing, right? But stakeholders take an odd viewpoint when it comes to “their” content. They will freely admit they don’t watch another website’s carousels, but they fully expect people to watch theirs!

    I’m not sure I agree that a carousel is a cheap way to do A/B testing. Because of the time delay between each image, it’s not a good A/B testing tool at all in my opinion. Better to do real testing on the site. However, I do see your point that there’s probably some stakeholders that think just because the first image had the most clicks that it’s the best image. [sigh] Clients! Right?

    In my opinion there is no need to artificially create an alternative for carousels. The alternative to carousels is to set up your home page in such a way that it passes the 5 second test. A good designer will use content and presentation best practices to successfully pass the 5 second test. I mentioned it in the article, but it bears repeating: To be successful, a home page MUST deliver the “Who you are, What you provide, and Why should visitors care” information in 5 seconds or less. It’s very easy to set up a 5 second test to determine if the home page is being successful (or not). In my opinion you don’t need technology solutions to pass the 5 second test, you just need clear, simple and helpful content, good navigation and satisfying design presentation on the home page. “Fixes” like carousels are just cop-outs for bad planning on the stakeholder’s and designer’s parts (sometimes more the former than later).

    Finally, I’m not aware of ANY websites that are not interested in conversion. Remember that ‘conversion’ doesn’t have to equal a sale of something on the site. It could be a B2B site that is generating leads, which implies that conversion is all about getting a prospect to download something, complete a form, or interact with an item on the site (hopefully leaving a bit of information that the B2B firm can use to contact the individual later). Even a library site has conversion. I can’t think of a client I’ve ever had that was not VERY interested in conversion. But then again perhaps somewhere there is a stakeholder that is not at all interested in improving business results via their website. I’ve just never met such a person!

    Finally, I’m not damning carousels just because of conversion data, I’m also damning them because frankly they just plain don’t work. Website analytics reports in terms of length of session data will most likely show that the vast majority of website visitors do not stay on the site for longer than 10 seconds. Unless your carousel displays all images in less than 10 seconds, the odds are the vast majority of website visitors are not there long enough to see all the carousel images, which means it’s pretty much a great big waste of time even creating one. Remember the golden rule of usability…



    Finally, and most importantly, the good news I believe is many designers today have already arrived at the same conclusion that carousels don’t work. The evidence is the ever increasing number of newly redesigned websites that are choosing to not use carousels. When I sampled 100 websites recently (a mix of B2B and B2C) I noted that less than 30% of them had carousels. In my opinion, it’s just a matter of time before carousels go the way of “website tunnels.” What’s a website ‘tunnel?’ It was a brief design style in the mid 1990s in which a ‘tunnel’ of several intermediary pages were first show to the user prior to the user actually getting to the website homepage. The flawed thinking at that time was by having a visitor go through this website tunnel, they would be so interested and curious about the homepage and website that they would stay on the site longer. “Tunnels” are long gone now, most designers probably never even heard of them, but in many ways the use of a Carousel recreated that idea that providing many alluring images to a visitor could get them to stick around longer. We now know, through both practice and data, that both Tunnels, and Carousels, are counter-productive to the true nature of website user experience, which is to help the visitor to find the information they need as quickly and efficiently as possible, while maintaining a satisfying design experience.

    PS – How do I know so much about website tunnels? Because back in the day (the mid 90s, and yes, I’m that old), I too tried them for a while, thinking that they must be a good thing if everyone else is using them! Ha! What an idiot I was. I eventually learned (the hard way) that tunnels were not at all a good idea. Data helped sway me then, and hopefully data about how bad carousels are today will help sway a few designers now.

  4. You’re very welcome Ilona. Hopefully this information will help you and others improve the user experience of websites! I appreciate the nice comment.

  5. Hey, thanks for your thoughtful response, Craig. We really are coming from same space, I can tell from your discussion of “tunnels” (and intercepts–I have a client right now who wants to strictly path audience personas into inane tunnels, in 2014! Egad. Can you say “abandonment”?).

    And I don’t disagree with you about the need for sites to convert, to economize and optimize clicks. I am driving that point home sometimes it seems more often than the SEO people do. In the attention economy, anything a click away has to EARN its engagement, and the clients I work (big established businesses) with never seem as hungry for clicks and engagement as lean and hard working little sites and startups. They think their site is like PE class, where you get clicks just for showing up.

    The part that I found limiting is the need to put the burden of conversion on the boundaries hero/carousel. Yes, the whole site should be about conversion. But isolating the hero/carousel to do that job just because it has the biggest real estate feels too much like the wrong-headed thinking that drove businesses to demand banner ad “click-thrus” in the late 90s. Brand awareness and messaging will also drive conversions, without onerous demands of a CLICK NOW button funnel within the carousel boundaries.

    I’ve never represented to a client that anything but the first slide will get attention. Anything beyond the first slide is gravy. The design pattern is simple, and users are by now well conditioned to manually stop a dot or indicator on a particular slide.

    But the bigger issue, the one you address in your reply above as well, is client ignorance about basic patterns of interactivity, heat and click maps, key performance indicators. We have a job to do to break through this, but in the meantime, what I think was my primary point above (sorry if it got lost in my general rant) holds:

    The comparison isn’t carousel or no carousel. This is sort of like the AOL busy signals on dialup that users got in the 90s. AOL knew they didn’t have enough lines, but they had no way of knowing HOW MANY people were lost to that busy signal.

    One can’t compare carousels against the invisible ghost users who bounce away at 10 seconds because with no carousel or A/B testing, the “placate the stakeholders” message will lay full claim to the hero spot, and it will NOT drive conversions.

    Yes, it is a noble project, to get a well-crafted and effective (and conversion-driving) message and image in the hero. Good luck with that. Something we can all strive for.

    Yes, wouldn’t it be nice if we could sell clients more often on A/B/C testing in that spot, with alternatives as big and as beautiful as the secondary slides in most carousels? But most client-approved A/B testing I’ve seen does little more than tweak syntax of a single “placate the stakeholders” message. Oh look, version A of this sentence works marginally better than version B of this sentence. They don’t test alt heros with the kind of variety of messaging you can find in a carousel.

    Like the AOL busy signal, demanding and placated stakeholders won’t ever know (and users won’t ever see) those secondary heros that might have been. They’ll just happily abandon the sites for oblivion.

    Don’t mind me. It’s been a cynical week! LOL. Thanks for the great discussion.

  6. Thanks Chris, I appreciate your thoughts on this. I think I’ll have to continue to disagree on the issue of carousel vs. no carousel comparison, but I absolutely agree with all your other points. Sadly, education of stakeholders seems to be a never ending job, but hopefully they will all read this post and discussion and instantly become educated! Well, we can hope, right?


Comments are closed.