(This is a posting from yesterday's keynote at SXSW. Please forgive the delay, the Internet connection here is iffy.)
Kathy Sierra asks an intriguing question at a technology convention: “Why are we all here?” If we can connect almost instantly via technology, then why do we bother travelling to be together in person?
The answer is simple and obvious. Computers, no matter how amazing, can't take the place of humans for one very important reason—they can not detect emotion. (Not to my fellow geeks: Sierra made a very funny point about how most of us tech folks can’t detect emotion either … Google Asperger’s Syndrome for fun).
The problem is that if we can't detect emotion, then we can't help users at the most crucial points of frustration. Sierra suggests a few ideas for solutions, for example, a “WTF button” or a graphically based entry into help files where users can click on the facial expression they are making at that moment.
The problem is so very apparent, Sierra says, when you look at FAQ’s and help files, which are so often stiff and linear. We need to give users understandable questions that don't assume prior knowledge, she said. Start at a higher level of dialog. Make tiny changes to your language, like using the word “you.”
It seems so simple, yet why do we so often get it wrong? Personally, I’d suggest that the formula of text=help is wrong.
When we write, we remove ourselves from our feelings. Socrates said, “The discovery of the alphabet will create forgetfulness in the learners’ souls, because they will not use their memories; they will trust the external written characters and not remember of themselves.” They will appear smart but will know nothing, he said.
Now, text can be used in interesting and interactive ways. Ford.com’s vehicle finder, which I worked on with a team at Avenue A | Razorfish uses a simply layover to help connect instructional text with the site.
But there must be other, more innovative ways to help users that are integrated and meaningful. There are ways to create interesting text based help files—that interact. What models can we draw on?
First, I would suggest that we all wikify our help files. Why can't users add or change help files to make them more useful based on their experiences? How about using audio that is navigable via verbal commands. Or what about video that users can annotate?
Those are just a few, very high level thoughts. I’d love to hear more!