Summary: Mobile phone users struggle mightily to use websites, even on high-end devices. To solve the problems, websites should provide special mobile versions.
I recently sat through many sessions in which usability test participants attempted to use websites on their mobile phones . What a cringeworthy experience — for both users and researchers. In terms of the user experience quality we observed, it was like stepping into a time machine for a quick trip back to 1998. The similarities were numerous:
- Abysmal success rates . I don't want to publish specific numbers until we've completed our next round of testing in London. But in the U.S. sessions, users failed more often than they succeeded when using their mobiles to perform tasks on websites.
- Download times dominate the user experience. Most pages take far too long to load, particularly on non-3G phones. But even the highest-end phones deliver much slower browsing than a desktop computer. As a result, users are reluctant to request additional pages and they easily give up.
- Scrolling causes major usability problems. In contrast to the 1990s, the problem is not that users don't scroll — it's that they scroll too much . On mobiles, they have to move their minuscule peephole back and forth so often that they lose track of both where they are and what's on the page. Often, they scroll right past something without noticing it. The effect of the reduced viewable area on users is strongly reminiscent of usability issues we found in tests with low-vision users . Using a mobile makes you a disabled user , and we all know that most sites ignore accessibility.
- Bloated pages hurt users. Most of the sites we studied wouldn't seem bloated on today's upsized PC monitors, but when rendered on a mobile they fairly explode with bloat. Users are frequently stumped by big images or by long pages that bury the items they want to see.
- Unfamiliarity with a browser's user interface limits the user's options. People use their devices suboptimally because they don't understand the UI. Desktop browsers are fairly stable, with few major changes between releases (tabbed browsing is probably the only one in the last decade). In contrast, many people get new mobile devices every two years, and the various models have vastly different browsing experiences. This fact also limits a user's ability to learn from observing friends and colleagues, who may have different phones.
- Reluctance to use websites on the mobile for many tasks, especially true of shopping. According to our participants, m-commerce has a dark future unless sites improve and earn users' trust.
- Search dominance . Okay, this is more prominent today than in 1998, but it was there, and it's certainly strong in mobile use.
- Old-media design . In the 1990s, many site designs mimicked good-looking print publications and offered weak interaction support. Today, sites are designed as, well, websites. More specifically, they're designed as desktop websites, and that's the wrong media form for mobile use; even on the best phones, driving the interaction is painful and simple designs are a must.
Usability Varies by Mobile Device Category
Our testing found 3 distinct classes of mobile user experience, which are mainly defined by screen size :
- Regular cellphones with a tiny screen . Often called feature phones , these devices account for the vast majority of the market (at least 85% in some statistics). They offer horrible usability , enabling only minimal interaction with websites.
- Smartphones , in a range of form factors, typically with a mid-sized screen and a full A-Z keypad. These devices sometimes feature 3G Internet connectivity and perhaps even WiFi. Smartphones offer bad usability , forcing users to struggle to complete website tasks.
- Full-screen phones (mainly the iPhone) with a nearly device-sized touchscreen and a true GUI driven by direct manipulation and touch gestures. These phones offer 3G Internet connectivity and even faster speeds when connecting through WiFi. They also offer impoverished usability ; only simple tasks are reasonably easy — and only then if users are on well-designed sites that are optimized for mobile.
In principle, the third category should include models beyond the iPhone, and we did observe a few users perform decently with other full-screen phones. In practice, however, most of the other full-screen devices we tested had usability that was so weak users didn't browse the Web much better with them than they did with regular smartphones. It's not enough for other vendors to copy the iPhone's hardware characteristics (a big touch screen). They also have to provide software with a high-usability GUI.
A Separate Mobile Site Is Best
For the best user performance, you should design different websites for each mobile device class — the smaller the screen, the fewer features, and the more scaled back your design. The very best option is to go beyond browsing and offer a specialized downloadable mobile application for your most devoted users. In practice, however, only the biggest and richest sites can afford all this extra work on top of their desktop-optimized website.
Moderately rich sites should build two mobile designs : one for low-end cellphones and another for smartphones and big-screen phones. This strategy is especially good if you're targeting a broad consumer audience with many feature-phone users. The small-phone experience is so different that it needs a dedicated and deeply scaled-back design, whereas the bigger phones benefit from a design that's mobile-friendly but not bare-bones. Feature-phone browsing is essentially a linear experience , whereas smartphone and full-screen browsing provide more of a GUI experience — albeit through a limited viewport.
For many sites, however, the only realistic option is to supplement the main site with a single mobile site , recognizing that it will serve plain cellphones poorly. This strategy often makes sense. After all, most low-end mobile users suffer such misery when they attempt to visit websites that they do so only for the most compelling tasks, and thus might not use your site anyway. So, if you have only one mobile site, target the medium-to-higher-end devices, as opposed to making a WAP-like site that everybody will hate.
Finally, not all sites need mobile versions . According to a diary study we conducted with users in 6 countries, people use their phones for a fairly narrow range of activities. So, because many mainstream websites won't see a lot of mobile users, they should just adapt their basic design to avoid the worst pitfalls for those few mobile users they'll get.
If your service makes sense for mobile users, offer at least one mobile-optimized design. Don't rely on "full-featured" browsers to display your main site , because doing so will cause endless usability problems. If your site has both a mobile design and a desktop design, serve the mobile version to all mobile users — even those with phones that support full-page browsing. (For users who need rare features that aren't in the mobile design, you should offer an easy way to switch to the full site.)
Half-Speed Progress, But Hope Ahead
Our last study of mobile usability was conducted in 2000, and my conclusion then was that mobile Web 2000 = desktop Web 1994 . Now, observing users accessing websites on their mobiles reminds me of testing wired users in 1998. In other words, during this 9-year period , we've seen 4 years worth of progress in mobile user experience. Roughly speaking, improvements in mobile usability moved at half the pace of wired usability .
So, why am I still bullish on mobile websites and online services?
First, it's not necessarily tragic to advance at half the rate of the mainstream Web during the second half of the '90s. That happened to be a period of explosive progress.
Second, mobile is the trend of the year in application design . While trends can be wrong, lots of interesting things are happening.
Third, we're turning a corner in mobile Web usability . Just as Apple's Macintosh heralded a breakthrough in personal computer usability 25 years ago, its iPhone is pioneering a similar breakthrough in mobile usability today.
The iPhone is certainly not perfect, and competitors could easily make better mobile devices. By "easily" I don't mean over a weekend. I simply mean that it's possible to do it given a strong focus on user experience and user-centered design; iPhone leaves a lot of ground for improvement. So far, however, iPhone competitors have been disappointing because they haven't been created with UCD.
Alan Kay famously said that the Mac was "the first computer worth criticizing." Similarly, the iPhone is the first mobile Internet device worth criticizing . It's a starting point for mobile online-services access, not an endpoint.
Although devices will get better, the big advances must come from websites . Sites (including intranets) must develop specialized designs that optimize the mobile user experience. Today, few sites have mobile versions, and those that do are usually very poorly designed, without knowledge of the special guidelines for mobile usability.
There is immense potential for advances in mobile usability as more website, intranet, and enterprise software designers build mobile versions and revamp their current designs for usability. The mainstream Web's state in 1998 actually provides a hopeful precedent: just a year later, in 1999, interest in Web usability began to explode as Internet managers realized how chasing "cool" rather than usable design yielded poor business results.
Let's hope history repeats itself for the mobile Web.