Recent Blog Posts

J.D. Power & Associates just released their 2005 U.S. Wireless Mobile Phone Evaluation Study(SM), which measures customer satisfaction among those who have owned their mobile phones for less than two years.

I've not been tracking this information for the past few years, and the last time I checked, battery life was still king. Now, however, there is an array of things that drive satisfaction in a phone. Top drivers are:

1. physical design (24%)
2. operation (22%)
3. features (20%)
4. handset durability (19%)
5. battery function (15%)

They report a major shift in importance of operation of the phone (up from 15% in 2003) and physical design of the handset (up from 19% in 2003).

Each of these features is a key driver to overall user experience, of course. The marketers more or less have control over the features of a phone; durability is driven by a combination of design, cost vs. features, and manufacturing; battery is driven by cost and engineering. The cost factors are interesting since the study reports that people are paying, on average, $10 less per phone since last year.

Physical design (particularly clamshell rather than candy bar, but including other factors) is driven by a combination of industrial and mechanical design along with marketing.

"Operation" is essentially the user interface (both design and implementation) of the phone - how users can get to the various features. This doesn't have any per-unit cost. Development investment in "operation" can be spread across multiple devices.

In effect, usability (ease of use) is driving 22% of customer satisfaction with phones. Let's not forget that satisfaction with a phone will increase the likelihood that the user stays with the same brand when making the next purchase, 18 months later.

I was recently told that a product design "didn't need usability work" because it was "just a reference design". Never mind the fact that in the reference design, developers and designers make almost irrevocable decisions about how the product will behave and look.

Perhaps there is an assumption that usability is still about screen design? Interaction (behavior) is at least as important, if not more so.

Or perhaps there is an assumption that a reference design will be used as a proof of concept, and not recommended code and design? History suggests nothing could be further from the truth.

Prototyping is an activity intended to show that a project could be done, and to learn about problems in the implementation and use. Product design is an activity intended to create an efficient, easy to use, easy to manufacture, easy to test, easy to expand market deliverable. Prototyping is used when the product direction is still unknown.

Sun's J2ME KVM (Kilobyte Virtual Machine, the environment in which MIDP applications run) reference designs have caused havoc with their technology adoption. Sun made the easy engineering and design decisions, making a prototype rather than a product. The companies implementing the KVMs then made even easier decisions: use Sun's code without truly understanding it.

The result is that J2ME applications load slowly and the user interface doesn't conform with Sun's recommendations, and applications are far less satisfying than they should be. Users might be willing to wait 30 seconds for a game to load, but not a real application. Developers write applications assuming the user interface works the way that Sun says, but the implementors have never even seen Sun's recommendations.

Please, folks: if you're going to make a reference implementation, spend the time and money to fully develop the product. You're making decisions that you and your users will have to live with for years to come.

The New York Times recently published an article on interruption science (registration required), which is particularly relevant to the future of mobile applications. The basic idea in this article is that people are spending less and less time working on problem solving ("actual work") and more dealing with interruptions - both in person, and electronic. The problem is that these interruptions are necessary to people's work, both as part of the job description and as social connections.

There are some potential solutions from the field of HCI. If you need to interrupt a user doing highly text-intensive work, a text interruption will not get the user's attention very well, but a sonic alert will completely draw the user's attention - so make sure you use sound for critical alerts.

Artificial intelligence techniques learn what computer activity means that you are working and focusing, and only alert you with incoming email messages or other information when it is critical. The amount of information available to do this in the mobile space can be much greater than the desktop space, as I discussed in my mobile context entry. As yet this is merely potential improvement. Note that Orative also addresses the issue of when to interrupt, although at a much lower leveal than sophisticated AI techniques. I expect that a hybrid of automatic techniques and manual techniques will become the dominant model for determining the user's current needs.

They have also found that using a 44 inch plasma display makes users 10% more efficient than with standard 17 inch displays. Obviously we do not want mobile phones with very large displays, and we immediately wonder what the efficiency cost is on a small mobile device display.

For now, few people are attempting to do complex, attention-intensive work on a mobile phone or PDA. Eventually this will change as a device somewhere between a PDA and a mobile PC becomes common. The mobile phone is far more likely to be a source of interruption. However, interruptions on mobile devices will be a much greater problem than they are in the desktop space, merely because the user is mobile and likely to be in a public space. Thus interruptions will arise from in-person colleagues, email messages, phone calls, and the non-office environment.

Further, this problem is going to be exacerbated by the small screen, which does not support mulitple windows to remind users of what they were doing. Current devices make no attempt to store what the user was just working on, and instead only give the user what they are currently doing. This pushes the responsibility for storing the entire set of currently-active tasks onto the user's short-term memory, which is woefully inappropriate for storing this information.

To make multiple-function devices practical work devices, we will need to adust the current interaction methods assumed in mobile device design. We must assume that the user is likely to be doing several things on the mobile device and several other things off of the mobile device (perhaps on other devices, perhaps elsewhere entirely), and aid the user in doing so. Let's get this memory load off the user while we sitll can.

I've talked to several professionals recently who claim that usability is irrelevent.

Excuse me?

When I ask further, I learn that their definition of the practice of usability is actually limited to reviewing products for ease of use, and making suggestions to change. Certainly the folks over at Usability Must Die appear to think so.

This definition of usability is gaining momentum. Certainly "Usability Specialist" jobs are generally all about testing, although I've found that "Usablity Engineer" positions are not. Then "user experience" specialists will write things like this: (source)

The idea of usability, sometimes known as 'human factors', existed long before the web. It involves observing users engaging in tasks and mediating between design and the end users' needs - ensuring that customers can achieve the original aims of the product, whatever that product may be. Today, everybody is into usability. Usability originated in ergonomics, information design and software design, and was epitomised in the 1960s and 1970s by military displays and cockpit panels. Now it has come of age. It used to focus on product design - for example, trying to make your video remote control work properly. Today, a combination of ubiquitous computing power and the internet has made internet usability appealing to just about every business worried that it might be out of touch.

In fact, human factors and ergonomics are largely synonomous. So says The Human Factors and Ergonomics Society, so also says Wikipedia. They each involve understanding human needs and characteristics and how to apply them to the design of products, furniture, airplanes, software, or anything else. There's a good reason why schools with these programs usually have degrees available in both engineering and psychology, with possible programs in computer science, technical commuication, and design.

The breakthrough with human factors (originating with US Air Force work, WWII) and ergonomics was simple: rather than requiring the human, the user, to adjust to the man-made environment, the environment (tools, doorways, cockpits) should be designed to "fit the task to the man".

The US tends to prefer "human factors" while the rest of the world tends to prefer "ergonomics"; there is also a vague connotation that "ergonomics" refers to neck-down design issues (think ergonomic chairs or mice) whereas "human factors" gets neck-up issues (cognition, emotion, etc.).

Usability is the degree to which a product is easy to use. It is one issue, one collection of measures, within the larger user experience. Pleasure, enjoyment, and learnability are other issues. Usability is measured in "usability testing", by asking users to try out the product, and observing them. However, usability is achieved by starting with good information about the users (using research, general principles, and so forth), applying good design techniques, and checking your work at key points using various usability testing techniques.

Note the critical role of design. I thought this role was so important that I picked up a minor in design when studying human factors and ergonomics.

There are times when usability is irrelevant, such as in game play. It might make sense to require the game player to achieve a degree of skill using an obscure interface before competitive scores can be achieved. Business needs must dictate the relative importance of usability for a certain product.

The main points I hear from usability foes is that enjoyment and usefulness are ignored by usability, and that usability is all about reviewing products. This simply isn't true. What is true is that usability is one of many measures for a product or service. Foes complain that Jakob Nielsen's site is ugly and unusable. Nielsen's site certainly would be improved with some decent visual design. In fact, I imagine that the architecture of the site could remain exactly the same, and that usability would improve if good visual design techniques were applied. Enjoyment or pleasure would definitely improve.

Oh, and what is the difference between usability testing and quality assurance? In theory, nothing. In practice, I have worked with QA teams repeatedly to try to get some well-crafted standards into the QA process, but I have yet to run into an organization that enables QA testers to do anything outside of a strict script. Only standards testable without users can be incorporated into the QA process.

I just learned about Orative, a software firm that uses some of the key components of chat systems like AIM, within the corporate firewall and for verbal communications.

Basically, they have software (J2ME, Symbian, BREW) that runs on handsets that lets users access the entire corporate phone directory, indicate status (in a meeting, unavailable, and so forth), and request callbacks indicating subject and priority. It is therefore an application with a text messaging component, used to enhance voice communications. Of course there is a server component as s well. The idea is to prevent voicemail tag, reduce meeting interruptions, and so forth.

This is a terrific idea. It combines many of the benefits of email (asynchronous messaging), chat (presence + synchronous messaging), and voice communications. I imagine that for people who spend most of their time talking to others within the company, it's a wonderful solution.

The big problem, in my opinion, is that I can't use it. I'm talking about small businesses, who spend most of their communications time with people outside the company. In my case, clients and partner organizations couldn't be included. Within larger companies, anybody who interfaces with the outside world would not be able to use the software.

So here's my recommendation: A business networking site like LinkedIn should install the server component on their site. People paying a subscription fee to use the voice service can allow anybody in their immediate contact list to see presence information for others in the list. Note the major revenue opportunity for LinkedIn or whomever sets up the external affinity groups.

We'll need to alter the software a little bit as well, because users can now be involved in multiple networks: one corporate, and any number of affinity networks. This would be an excellent tool for those involved in any flavor of business development or partnership work.

This application (even without my recommended enhancement) is a wonderful example of an application that is starting to achieve some of the promise of mobile: It shares information with the device's phone book, adds an application running on the device with ongoing data connectivity, connects to a central server, is massively networked within a community, and also enhances voice communication. This promise of mobile is what I believe will make mobile the dominant application environment in the next 10 years, if we can just get some application providers to start making multi-context applications. Good job, Orative.

While the PalmOS future appears to be bleak, Palm itself is doing something interesting with its adoption of Windows Mobile for a version of the next Treo, with Symbian UIQ likely to come next. I was especially heartened by Microsoft's comments that Palm negotiated several changes to software and requirements to make their device look like a Palm. Of course, we should expect Microsoft to incorporate any transferrable Palm ideas and putting them in the next set of device requirements and software.

Honestly, I expect the device will look like it came from the Handspring design team, not the Palm design team (within the larger constraints of Windows Mobile SmartPhone). Palm devices (i.e., those by Palm outside the Treo line) have been somewhat uninspiring for the past several years. My LifeDrive, for example, has many of the same physical characteristics as the Treo 600/650, but the software is not well integrated for thumb or 5-way rocker use. They took some of the lessons from the Treo but not all (hey, isn't that what Microsoft does?)

So, based on recent reports, it looks like Palm's core competency, the design of mobile devices' user experience, is being preserved. This is good since they had to purchase it back from Handspring. They don't appear to have the ability to build quality hardware, particularly the radio components, if recent quality reports and lawsuits have any merit. Certainly the radio on my Treo failed.

So what is the future of Palm?

As an independent device manufacturer with no inside track for any operating system, their weaknesses in marketing and quality spell a further downward spiral. After all, other companies can and do make devices with good user experience. Their divorce from PalmOS the company is likely to cause that OS to become less relevant in more developed markets, so they don't have an inside track there.

As an acquisition target, however, Palm is very interesting. What if they found a company who could make and distribute smartphones, had good relationships with the carriers, but who didn't have a (Western) brand identity? I'm talking particularly about the Asian device manufacturers, particularly Samsung, Sanyo, and LG. Samsung, for example, will make a device with any operating system that's out there, and does not have a strong design identity. LG is struggling to get out of Samsung's shadow. Sanyo makes very high quality devices, but is too accustomed to the Japanese strong-carrier business model to have much of an identity.

Palm's strengths could complement the strengths of one of these Asian manufacturers.