Web applications still put too much strain on the user as they need to remember complicated preferences, codes and you fill in forms the same over and over again. In order to find out how a web application works, you typically just muddle through the interface. Though most web applications require you to understand the jargon they use and how previous stuff you put in affects the system as a whole when you're still a beginner.
The general assumption is that people will find their own solutions and conventions within your application. But considerate web applications try to do the thinking for the user, users shouldn't have to do the hard work. Web applications should notice context. Is the user is new to the web application, then give this user an introduction on how the system would look if it was populated and information and how to achieve basic tasks, a term used for this is zero-content.
As people are becoming increasingly experienced with building interactions for the first-time user, intermediates get left behind. - Me
When I am trying to categorize my content with 100 pages already the interaction is the same as setting up my categories with 0 pages. Context is becoming increasingly important to solve usability problems, as web applications usually can't serve secondary audiences of their application without sacrificing parts the primary audience's user experience. Quite often, audiences of web applications differ considerably in expectations and current knowledge of the domain.
Adaptive interfaces Part 1 explained that interfaces should change upon unique user behaviour. So tasks that you repeat often in a similar way will make obvious patterns. But tasks that you don't preform often, and are highly dependant on the context, tend to make a lot of vague patterns (many assumptions to be made). Developers usually back off here, since wrong assumptions kills user experiences. However what if you would ask the question: What assumptions can I apply to filter out functionality that is highly unlikely. This way the system can learn over time if it makes the right assumptions.
Decision-set reduction
Alan Cooper talks about Decision-set reduction that users unconsciously reduce a large amount of choices to a small amount and go with that selection in a similar fashion every time. People who set up multiple blogs, CMSs or wikis will recognize this. Most of the time you go through the same kind of process setting your preferences in the web application (ie. setting up plugins). It's like choosing that Coca-Cola is your favourite drink, once upon a time from the infinite amount of drinks you choose cola or a couple more to drink and from that point most of the time you drink those.
Web Applications usually have huge amounts of data in their databases they don't use. The key is tracking interactions over time—do patterns emerge? Then, you have a interaction that you can easily enhance (take away choices - offer relevant choices). Interactions that show a lot of vague patterns can be tackled by making a better distinction between the types of user behaviour.
Usability problems scale
Web Applications have many different types of users, yet most only offer one interface. In a lot of cases, you are sacrificing user experience for a potentially large number of users, to support the majority.
Facebook, making any change to their site, instantly affects eight million people. If even one percent has issues with the change, that’s 80,000 affected users. - Jared Spool
You can now focus on a certain type of user who is having a problem or is in a certain context without putting strain on the majority of users. Adaptive interfaces can also counter the almost unavoidable process of a cluttered interface over time, even 37Signals is suffering from feature creep. The feature creep is avoided by adapting the new features only to users who actually need it and not of the bat introducing them to new users. This is, however, a completely different approach to software development, as most of the features will be introduced silently.
Related link
UIE : Five Usability Challenges of Web-Based Application
Discussion
Really well written. I kind of like 37 signals approach, of minimum features, rather than focus on doing 5000 things, so-so well, focus on 50-100 really well.
But I think a key question, is how to determine what an end-user really wants in design, useage and programming for a web-application, when most of the time, there are many hidden things, they either do not explain, or can not show or explain.
How do you approach that?
June 4, 2008
You approach this by creating a mental model of the user. Usually people step into the developers process trying to explain and form their answers towards how it could be implemented (pleasing the one asking the questions). With mental models you try to create a mental representation of how the user expects to achieve a goal. With this you try to visualise how people behave and what underlying decisions and emotion they have against how you support those needs.
Almost rarely do users explain or show behaviour and emotion consciously but they do express it. User-interviews (not surveys!) supply more context, as you engage in an conversation you can ask follow up questions. When you do these user-interviews usually you can refine traditional approaches such as persona's, for adaptive interfaces this is important as you can create a deeper understanding of the different types of users and why their behaviour is different.
Indi Young
http://www.rosenfeldmedia.com/books/mental-models/
June 4, 2008
Then after a web-application is successfully released, do you go through any kind of review process, to see if the user experience is as expected, and what to plan, to revise for next version?
Although I much prefer to avoid revisions, other than new unexpected features. I strongly prefer to get it right the first time, but that is so rare, in my experiences.
June 4, 2008
Stepping outside a traditional software mindset is important, you have to test often and revise a lot. But over time your web application should show signs of meeting user experiences better. Usability is not like a project, that you do and your just finished with it until the next version comes. Its a repetitive process that should change the interface over time instead of abruptly every major release.
So your review process would be continuously testing the interface and implanting these revisions on the go, so next time you test it you can see if its become more usable and more pleasant to use.
June 5, 2008
Post new comment