by Chauncey Wilson on July 20, 2009
User Experience and usability practitioners are on a continuous hunt for problems that plague our users. This seems straightforward – find problems from testing, user forums, observation, and other methods, prioritize the problems, and generate solutions that eliminate the complaint. However, some events that we call problems in one context may not be problems in another. Take a simple example of a drop-down list that allows you to choose a single item from a longer list. This control (assuming that it is well labeled and you can press a letter to jump to the first instance of that letter in the list), may not present a problem with 5, 50, or even 500 items; but what if you have 50,000 items with similar beginnings – then you have a scalability problem. Some user interface controls work well until you hit a particular threshold – here more than 500 items – and then suddenly, a simple act of choosing an item becomes an irritating problem.
In complex products, it is sometimes hard to know when a problem begins. Some problems might begin weeks before you actually encounter any negative event. An example of this might be a choice in a properties dialog that you set during installation that you forgot about and then weeks later, when you are working against a deadline, doing something simple with your tool, BAM! CRASH! BOOM! Your data are destroyed and you have no backup. You talk about a crash as the problem, but it was really the result of an isolated change made earlier.
One of the conundrums of user experience design is that we have to design systems that accommodate different user groups, from first-time users to gurus. A system that delights the first-time user by guiding her through tasks (high learnability) might frustrate the guru who expects shortcuts, ways to combine functions through macros, or a command line. So user experience can influence whether a particular event is a problem or not.
Age makes a difference in determining whether a design feature is a problem or not. I recently turned 59 and find it almost impossible to read the registration codes on those plastic containers for software DVDs. Colleagues in their 20s and 30s may not find the tiny 7 point text a problem, but I surely do (and carry a magnifying glass around in my computer bag).
What other contextual issues affect whether something is perceived as a problem? Weather can affect whether something is a problem. My iPhone worked well for the first 5 months, but then one day I wore gloves and discovered that trying to dial a number with gloves didn’t work with the capacitive touch screen. Apple has filed a patent for gloves that transmit a charge in cold weather to mitigate this problem (see http://www.appleinsider.com/articles/09/01/01/apple_files_for_patent_on_winter_friendly_iphone_gloves.html ) . Years ago, I worked on an early computer system that police could use while chasing evil doers. The system worked well in static testing, but when the laptop was used in simulated high speed chases on bumpy roads, it was a disaster. The officer trying to operate the system had trouble typing in the license plate and reading the output (THIS MAN IS DANGEROUS AND HAS POWERFUL WEAPONS).
So, what constitutes a real usability problem is often dependent on context and perspective and what is a real problem to one user might be a delighter to another. Some questions that you might consider to understand (and anticipate) the impact of context on how events in user research are labeled include:
- What is the range of environments in which your product will be used?
- How does performance interact with usability?
- What activities become difficult when you scale something up, like the number of collaborators working with the same data? Something that works well with a few people may become a monster when you involve dozens of people working on the same thing.
- How do your data samples affect whether tasks are easy or hard? Do you use large samples with dirty data (the real world) or toy samples with clean data?
- Did you consider the breadth of users in your research? Did you focus too much on learnability to make your novice users happy, only to irritate your expert users?
For some other thoughts on this topic, you can read a related article that I wrote in 2007 for the ACM magazine, interactions:
Wilson, C. E. 2007. The problem with usability problems: context is critical. interactions 14, 5 (Sep. 2007), 46-ff.