By now everyone knows that page load time is critical – It’s the one unifying factor that can ruin absolutely any app or website in existence. As optimization pushes the limits on how fast content can be delivered, milliseconds can make a huge difference to a user, as well as a company’s bottom line. Interestingly, it turns out there’s something that could be even more important than how long it actually takes – and that’s how long a user thinks it takes. At the past few Velocity conferences, this topic has gained momentous attention and acknowledgement, and with good reason.
The numbers aren’t the only thing that matters.
I think it’s safe to say that most users care less about the site statistics and more about the experience of the website interaction. Of course these two are related, but a responsive user experience beats a glamorous one. In a recent post on website response times, usability guru Jakob Nielsen points to two psychological reasons that people crave this responsive interaction:
- People like to feel in control. The more they feel like they are in control of the system and the more they will like it.
- People have short term memory. A long page load time could not only make users lose their patience, but also their focus. After about 10 seconds, people start to have difficulty keeping their brains on track. A 1 second response time seems to be about fast enough to keep user’s flow of thought seamless.
When users are able to truly move freely and engage more with a site, they are able to focus on the content. When a website feels unresponsive or slow, it distracts the users from the content.
The Placebo Effect
The time to consider the human aspect of performance engineering has come. Using techniques that distract the user, or hide the wait, can trick users into perceiving the site as faster. Now, I don’t mean “trick” in any devious, sleight of hand kind of way. I just want people to feel happier, less stressed, more productive. This is why there are TV’s in waiting rooms, mirrors in elevators, and why the Disneyland lines are always hidden behind walls or buildings.
How can we measure what people perceive to be finished?
Coming up with a universal algorithm that can be applied to any site and consistently measure perceived performance is a tremendous undertaking. As of now, I couldn’t find many tools for measuring perceived load time. Webpagetest offers a Speed Index, which measures how quickly the page contents are visually populated, although there are some limitations.
Chao Feng, a performance engineer at Microsoft is developing an exciting tool to measure Page Phase Time (PPT), which measures performance by analyzing pixel changing velocity and acceleration. You can check out the basis of the algorithms here.
How can we optimize perceived performance?
The big question is, how can we optimize how fast a user considers a page “done”? Although there is tons of room for research and further exploration in this area, some interesting tricks that I’ve found include:
- Make the busy indicator feel faster. The busy indicator (think rainbow wheel, loading status bar, etc) draws attention to the fact that users are waiting. Which is why some people have made the somewhat controversial suggestion to remove the busy indicator altogether. Uxmovement.com gives great suggestions on this, such as using backwards facing ribbons in a progress bar.
- Load the main part of the page first. Our eyes are naturally drawn to the center of the page. It’s more important that the main content the user is looking for is loaded than the“About the Blogger”part of the page (Although as a blogger, I find this offensive.).
- Make buttons or hyperlinks give feedback quickly. This could be as simple as a color change. This indicates to the user that the action they took is already happening, although the overall speed of the page load might actually be the same. People like to know that the button they clicked on was pressed!
- Consider how fast you clear the page before loading the next one. There’s some debate here. Steve Souders argues that waiting to clear the page until just before the new page is ready will make it appear faster. Check out his blog post on the Perception of Speed here to test it out for yourself!
Overall, although perceived load time is something that will vary from person to person, it is definitely something worth considering in the development of your site. This is just one more way to make the user experience as pleasing as possible. A good experience means efficiency, conversion rates and customer loyalty.
Your turn. How do you define your user experience with a website? Do you have a different idea on how to improve it?