Do slideshows work?

, Wednesday, January 19th, 2011

Homepage slideshows are ubiquitous. You know what I mean: the rotating, clickable content blocks that increasingly get prime real estate on the homepage. They are stylish and flashy (whether or not they’re made in Flash), and you can find them on lots and lots of nonprofit websites – including many that we’ve built. But as Tim alluded, just because they’re pretty doesn’t mean they’re always necessary, or even a good idea.

We use analytics tracking on the homepage slideshows for a number of our clients.  I got curious about the results, one day, while I was helping one client look for the answer to a particular question: why did their new report not get the traffic they were hoping for? They were promoting it on their slideshow, so I started there.  What I saw was so striking (if not necessarily shocking) that I broadened my scope.  I looked at all the slideshows where we had analytics data, and asked: do these things really work?

“Work” is a relative term.  It depends on the goal you’re trying to meet.  With an old-fashioned slideshow, the kind where you show off your 300 best vacation photos to all of your friends, you’d probably call it a success if your audience makes lots of appreciative noises, and a failure if they fall asleep.

We like setting goals, so we have a pretty good sense of the needs that slideshows are meant to fill. Most often, they’re meant to:

  • Promote important content and announcements
  • “Engage” visitors
  • Show off compelling, mission-related visuals
  • Establish a personality or aesthetic for the site
  • Make it easier to manage what goes on the homepage

The last goal is interesting to me. Slideshows are a very easy way to manage content. Want something on the homepage? Give me a slide, and it’ll be the first thing visitors see on the homepage… until someone else comes along, and puts another slide on top of yours. (We’ve always insisted on a limit on the number of slides – usually 5, occasionally as many as 8 – so eventually each slide will rotate off.)

Slideshows make it easy for web content managers to settle the all-important issue of “I need space on the homepage” without hurt feelings or bad politics. Back in the day, the common solution was to have really long, scrolling homepages. Today, the homepage looks cleaner – the extraneous content is buried in the slideshow or other expanding clickamajigs, and displayed one pretty piece at a time. Certainly, this is easier for the web manager, so this goal is a success.

But what about other goals?

Most of the goals on this list are subjective, and hard to measure quantitatively. But the first two, promotion of content and visitor engagement, are things we can concretely measure with analytics. I figured that, if slideshows were successfully promoting content and drawing users in, you should see lots of people clicking through the slides for more information.

That’s not what I saw.

Granted, this is not a scientific study. For hard proof, I’d love to have some usability testing data to back up the analytics. However, this was data from four dissimilar nonprofits over a period of several months – and the same patterns held for each.

As slideshows get longer, click-through rate falls off dramatically

In each case, one slide would get a nontrivial number of clicks. Usually, this was the first slide, and it averaged clicks from about 1% of homepage views. The next slide would get about half as many.  The next slide, half again. Beyond that, the click rate rapidly dwindled towards zero.  I saw this pattern in every case I looked at.

The takeaway: people might notice the first couple slides, but they’re not likely to click, and beyond that, no one is even seeing them.

It’s easy to guess why this would happen. Most visitors spend only a few seconds on the homepage before deciding where to go next (or leaving). If your information architecture is good, they won’t need to wait around and see your slides. I would also guess that some visitors don’t understand that slideshows are clickable, or get frustrated when it changes before they can click – though, again, usability testing could confirm that.

A poorly designed slideshow can exacerbate the problem.  Timing the transitions is tricky – too fast, and the slide will pass before anyone can click. Too slow, and you’ll only ever see one slide.  If you do miss a slide that’s interesting, the controls are often so hidden or unintuitive that it’s hard to get back – you just have to wait for it to come around again… if you don’t lose patience and go somewhere else.

It seems that a homepage slideshow is not the best way to promote critical content. It may help set a tone for your site, and certainly makes it easier to slot content onto the homepage.  But it’s probably not a good way to encourage users to click.

If you’ve got a slideshow on your homepage, is it meeting your goals? There’s one easy way to find out – install some extra analytics tracking on your own slideshow, and see just how it’s doing.

3 Responses to “Do slideshows work?”

  1. AlexJB Says:

    We re-designed our site last year and introduced the ‘hero bar’ slideshow.

    My take on the GA metrics I’ve seen so far is that clickthrough is indeed quite poor. Definitely lower than we had hoped. No regrets on our design, as it’s definitely sexier and more professional looking than what we had before, but the poor traction on the heros is a bit of a bummer, as it does take a bit more work to produce them consistently.

    On the previous design, we had a couple of ‘promo’ areas on the homepage, and one of these days I’m going to carve out enough time to compare the clickthrough on that with our heros. Arguably not apples to apples, but pretty close.

  2. Tweets that mention Do slideshows work? | Beaconfire Wire -- Topsy.com Says:

    [...] This post was mentioned on Twitter by Rob Pierson, Jo Miles, Beaconfire, Tony Kopetchny, Chris vonSpiegelfeld and others. Chris vonSpiegelfeld said: Homepage slideshow efficacy http://ow.ly/3GN2G – Pretty considerable drop off after first slide #uxd [...]

  3. Links in #nptech – 2/13/11 Says:

    [...] Do slideshows work? [...]