How Can the Peak-End Effect Help Users Get the Best out of Any Experience?
Hiroto KakuIn April, I attended an interesting presentation at the CHI conference in Seoul, South Korea (you can read about my overall experience at CHI here) called, Examining the Peak-End Effects of Subjective Experience.
So first of all, what exactly is the peak-end effect? When we experience something, whether it's buying something on a website, or eating at a restaurant, the moment-to-moment experience is not constant throughout. Some parts tend to be better or worse than other parts, and when the user takes into account the entirety of these experiences when they think about re-visiting that experience. The peak effect states that intense positive or negative moments will be weighed more heavily when reviewing the overall experience. In addition, due to the end effect, the user will be heavily influenced by how the experience ended when rating it. These two effects combined, the peak-end effect, can have a large impact on how the user feels upon looking back at an experience.
I'm sure it's something a lot of us have experienced in our daily lives. Take the amusement park, for example. Wait times for rides can sometimes be over an hour. Even though the ride itself may be 10 minutes or so long, the experience it can give us can be so intense, that it may make up for the entire length of the wait, leaving you with a positive overall experience. The effect can work negatively as well. A single negative peak--whether it is a movie plot, eating at a restaurant, etc.--can leave a bad impression, no matter how the rest of the experience went.
From the presentation, we also heard of another example where researchers from Carnegie Mellon University had users compare progress bars that advanced at different rates. Some would load quickly at first, but slow down near the end; others advanced at a constant rate; and some loaded slowly at first but sped up near the end. Users preferred this final progress bar the most, even though the different types of progress bars that they saw all took the same 5 seconds to complete. Here, the end effect had left enough of a positive impression for users to rate that particular progress bar higher than the others.
In the presentation I attended, the lecturer spoke about his experiment that required users to perform tasks with identical objectives: setting 25 digital sliders on a computer screen to designated levels. The sliders were spread over a series of five pages with the number of sliders on each page being varied. Some pages had only 2 sliders and were easy to complete, while other pages felt more tedious to complete, having up to 7 sliders. These were distributed so that each user would experience an easy peak or difficult peak during their task, and could end on an easy page or difficult page. The researcher then asked users which of the sequences they preferred. The results showed that users slightly preferred tasks with just easy peaks or just easy endings, but a large difference in preference was only seen when the task had both an easy peak and an easy ending.
As UX practitioners, how can we leverage these results for improving products and services? Let's take websites for example. Unfortunately, not all online tasks are enjoyable; we need to fill in long forms, sites may be slow to load, and things may not work as expected. Hopefully, we can mitigate these pain points, but when it's inevitable, it's important to remember about the peak-end effect. A positive peak experience and a nice ending can leave a positive overall impression, and hopefully, keep the user coming back.