Interfaces with Good Aftertastes: Hacking People’s Memory
Here’s the experiment:
Sit someone down in your best headphones and play them (A) 4 seconds of horrendous and loud noise, and (B) The exact same thing but followed by the same noise just not as loud. They’ve now listened to two rather discomforting segments, one of which was 4 seconds long, and the other of which was 8 seconds long.
Now ask them which of the two sounds they’d rather hear again. Rationally, they’d choose A — choice B is the same amount of discomfort plus some extra discomfort thrown in for free. It’s a no-brainer. On average, though, people will opt to listen to noise B. Even though it’s objectively worse.
If you’ve ever wanted to prove your significant other is acting irrationally, you’ve now got the experiment to show it.
Why do people choose an objectively worse option? Because we view our experiences through the subjective lens of memory. The noise may be objectively worse, but it’s subjectively better.
To understand why that’s the case, we have to look at how memory works. There’s just too much information to remember about the world we experience, and so to compensate our memory compresses memories down to their essentials. It throws almost all of our experiences and sensory perceptions away. Sometimes our brain doesn’t choose the “right” essentials to keep. Remembering is fundamentally lossy. And that can be exploited.
The two most important factors that influences how much we remember liking an experience are (1) it’s largest extreme and (2) how it ends. It’s called the peak-end algorithm. It’s why if a concert gets off to a rocky twenty-minute start but ends strong you’ll leave happy, whereas if it starts strong but has a bad final ten minutes you’ll leave disappointed.
Back to the horrid-sound-through-the-headphone example. Both sounds have the same extremeness, but even thought the second is longer, it ends less gratingly. Thus, people prefer it. It’s a real-life demonstration of hacking the way our memory works.
That people’s preferences are dictated by the peak-end nature of memory has interface ramifications. Take the lowly and ubiquitous progress bar.
Naively, you’d expect it not to matter how the indicator proceeds: you have to wait the same amount to matter how it scrolls across the screen. In tests, it turns out that perceived speed and user satisfaction both go way up if the progress bar moves quickly at the end. When asked why, participants described feeling cheated when the progress bar started quickly and ended slowly, and that the program “knew what it was doing” when it front-loaded all the hard work and then went quickly at the end. Knowing that memory remembers via the peak-end algorithm, however, means we know why people greatly prefer progress bars that accelerate towards a finish: despite the same waiting time, the ending is better. It’s the aftertaste that matters.
A little bit of cognitive psychology goes a long way in interface design. Try to make your interfaces have a good aftertaste. You’ll have happier users. And have the satisfaction of knowing you hacked their brain.
Question: What’s your favorite specific interfaces (like the progress bar) can be improved by sweetening its aftertaste? One of mine is making loading throbbers/spinners spin faster over time. That way, if the user is waiting for a while, it feels like the computer is working over-time to finish whatever it’s doing.
No related posts.