Summary: The current generation of mobile Internet products and services has miserable usability (as shown at the DEMOmobile 2000 conference). New devices like Blackberry, Modo, and a prototype Microsoft telephone do better.
The DEMOmobile 2000 conference in early September 2000 clearly showed that mobile Internet is still not ready for real use. Almost all of the demos were complete flops when viewed with a skeptical usability-trained eye. Sure, the presenter was capable of laboriously walking through the necessary steps, but an average user having to perform a real task would not be so lucky.
For example, one of the demos concerned use of a WAP phone to order supplies for a construction project. It does make sense for a foreman to use a mobile device in the field, though it might be better to use a bigger, more supportive device in the car or in the shack. But then the demo proceeded to let the "purchasing manager" pull up the order on another cellphone to approve it. Come on, wouldn't the purchasing manager sit at a comfortable computer in the office with other necessary records within easy reach?
Worst of all, the demo only proceeded because neither the "foreman" nor the "purchasing manager" had any problems or issues with the order. They just clicked the "buy this" button. In real life, there are always exceptions. One of the first things you learn about task analysis is that no case follows the average rules to the letter. Any amount of real-world complexity and the demo would have failed.
Several services were shown in both WAP and Palm Pilot versions: every time, it was clear that the deck-of-cards form factor (of the Palm) offered immensely superior usability relative to anything based on a telephone. The bigger screen allows for the display of enough information to be useful. But even more important, the direct interaction with user interface widgets by touching/moving a pen to the screen provides a much more natural feel than the indirect "scroll-a-wheel" interface on the phones. As we found in a recent field study in London, WAP almost always has low usability.
New and better devices are coming. Informal discussions with conference attendees showed that Blackberry is taking off as the best and most loved product for mobile connectivity. A huge number of people had Blackberries (of course, most attendees at an expensive conference like Demo can afford $60 per month to read their email when out of the office — can average users? not so sure).
Pocket PC: Mobile Companion to Outlook
The most interesting presentation at DEMOmobile'2000 was by Microsoft. First, the presenter agreed with my long-held position that Windows CE does not work. Microsoft is notorious for not discussing the weaknesses of a product until they are shipping the successor, but at least the presenter did admit that it had been a mistake to take a user interface designed for a big screen and shoehorn it onto a small screen. The new Pocket PC design does seem better than CE.
Rather than making a miniature PC (as implied by the name), the design goal for Pocket PC was to create a mobile companion to Outlook<. A great strategy since I believe that most users will continue to have a base station in the form of a full-screen device in their home or office. It is Microsoft's job to make sure that this full-screen base continues to run Outlook as its primary software. It is everybody else's job to create a better communications central, but so far I am not holding my breath for anything to replace Outlook.
Kill the Telephone Keypad
Microsoft also demoed the best telephone at the conference: a cellphone with a high-resolution color display that could display standard HTML content (running Pocket Internet Explorer on the phone). It is interesting to see Microsoft work on hardware innovations, even though I think it's a doomed strategy to build mobile Internet devices as mutated telephones. Much better to build a mobile communicator and allow it to include voice as one of its datatypes.
The prototype Microsoft telephone had the best quality screen I have seen yet on a telephone, but still wasted more than a third of its physical surface on a numeric keypad. Let's just get rid of the keys and spend every available square millimeter on pixels.
Most of the telephone numbers you call come from one of these sources:
- your address book
- an email message (or equivalent such as SMS or paging messages)
- online lookup in yellow pages, telephone directories, corporate websites
Instead of keying in ten digits (a horrible UI if ever there was one), it is better to use the screen and simply tap the name of the person or company you want to call.
Update added 2007:
Apple's new iPhone finally implements my recommendation from 2000 to make a mobile device that spends its entire surface on a screen and doesn't have the traditional push-buttons. It only took 7 years for somebody to try, but the other phone vendors can't say that they weren't warned.
Modo: One-Handed Minimal UI
The most interesting new device shown at DEMOmobile'2000 was the Modo. I ought to have discussed it at the top of this article, following the content usability rule of beginning with the most important information. However, I had a conflicting usability goal in deferring this section: hoping that the photo would have finished downloading by the time you have scrolled down this far.
Modo device displaying the main menu ring.
Modo's industrial design reminds me of the vision concepts for future pagers shown by Philips in 1996. There is something about a plastic egg that feels more interesting and personal than the square forms that dominate consumer electronics.
Modo is operated by a single hand:
- The user's index finger rests on the back button (always the most important feature in any navigation interface)
- The user's thumb rests on a wheel with two functions:
- Scrolling the wheel moves the selection up and down the screen (and also scrolls the text when hitting the bottom or top of the screen)
- Pressing the wheel activates the current selection (which usually implies following a hypertext link)
- Modo also has an on/off button and a button to activate back-lighting of the display, but they are not touched during normal use
This description assumes right-handed use. From the company website, it does not seem that there is a different version for lefties.
One-handed operation is great for a mobile device. You often need to use your second hand for carrying your briefcase, holding on to a strap on the bus, or some other purpose that makes two-handed use less convenient than it is in an office setting.
Modo has a very scaled-back user interface. It is a true information appliance that only does a single thing: providing entertainment and "going out" listings for the city you are in. There is no configuration. If you bring your Modo on a trip to a new city, it automatically downloads the information about the new city from the transmissions it is receiving from the local network.
There is also no payment or registration interface. The device costs a flat fee of $99 up front; the ongoing information service is free. The hope is to fund the ongoing service out of advertising revenues. Interesting business model which I don't think will work:
- Advertising does not work on traditional websites, where you have a big color screen and the ability to follow a link to an advertiser's fully-featured site. Mobile use is much more time-critical and provides a much more scaled-back environment with less room for ads and less willingness on the part of the users to tolerate interruptions.
- An advisory service like Modo depends completely on trust: if users feel that advertisers get better ratings or preferred placement, they will stop relying on it. Even if advertisers are treated the same as everybody else and get bad reviews when they deserve it, users may think that the service is biased since the old "church-state separation" is not well understood in new media.
In reviewing early content on prototype Modos, it is clear that usability will be impacted by very minute details in the way the text is written. For example, the line feeds were often wrong and reduced the readability of the text. Considering that all Modo content is written explicitly for the device and its tiny screen, it will be important for the editors to pay closer attention to word-wrapping and line feeds.
I previously coined the term microcontent to refer to Web design elements like headlines and page titles. Modo highlights the need for even more fine-grained content usability, which I will have to call nanocontent. On mobile devices, copyfitting has to be done down to the character to ensure, for example, that the most information-carrying part of a headline is represented in the first 18 characters on certain WAP phones. Or to ensure that lines break in ways that maximizes the readability of the content.
Update Added October 26, 2000: Modo Dead
Unfortunately, my prediction turned out to be correct: Modo's business model did not work, so the company went out of business in late October 2000. Too bad, because I really did like the design.
- Report from DEMOmobile in the Los Angeles Times (Sept. 11, 2000)
- WAP, Europe's Wireless Dud?, Washington Post (Sept. 15, 2000), quoting the head of Internet operations for the largest bank in Sweden: "We have provided online banking via WAP to our customers since early this year. [...] You go down, down, down all these menus, and you wait, wait, wait each time. You're straining to read text on this tiny screen on the phone. Eventually, people just give up."
- Reader comments on this Alertbox (including whether left-handed users are better off with the current Modo and whether phone/PDA convergence is a good idea).
- My report on the field study of WAP usability we ran in London in the fall of 2000.