information architecture
Oct 26, 2005, 03:00am EDT
In a former life, I worked on a project for providing information during a crisis situation. The basic concept was pretty simple: during an emergency (like an earthquake or a hurricane), load a bunch of PDAs with local addresses and phone numbers (like, where the pharmacies in the city are or phone numbers of taxi companies) and let volunteer workers use the PDAs as they’re out helping victims of the emergency. Since the volunteer might not be local to the area, the PDA helps her/him find local resources.
There was a central database that had information broken down by categories (pharmacies, transportation, shelters, etc) and geographic regions so when an emergency occurred you’d just sync the PDAs with information for the region.
The part I worked on was the web site which was used to add information to the database. Our plan was to have volunteers enter data manually via a web interface. Aside from being the first real web application I worked on, it was also my first eye opening experience with usability.
After I had an initial web interface working, I let the nurses (i.e., not exceptionally tech savvy, but our actual end users) we were collaborating with enter some data into the system. As I quickly learned, they had a hard time entering data into the right categories. Most of the mis-categorizations were in similar categories, but not quite right. For example, there were entries for taxi cab companies in the ‘Bus’ category. The reason for the difficulty was that before you could add an entry into the right category, you had to create the category first. For me, the programmer, this seemed obvious. The database had a Category table and each data entry had a foreign key into the Category table; thus you had to first create a category (by selecting the “Add New Category” link in the side bar) before creating the entry (the “Add New Data” link in the side bar). It looked something like this:
As it turned out, this was not obvious to people entering data. Their task was to enter information about a taxi company so they would go to “Add New Data” first. When they went to pick a category for the entry, the selected from the drop down menu of existing categories. If an exact category wasn’t found, they picked whatever was closest.
So the obvious lesson I learned from this was to allow people to create categories on the fly as they’re creating an entry. This also removed the need for the “Add New Category” link on the side bar.
In hindsight, this is all obvious. I should have started by focusing on the task rather than building the UI to reflect the database schema which none of the users care about. But I bring this up now, because I see the exact same problem in wordpress.com‘s Links management page.
Links can only go in existing categories and if you want to put them in a new category, you need to know to select “Link Categories” and create the category before you can add the link. This is a general feeling I get when I use wordpress. I feel like the UI was designed to reflect the database schema rather than user tasks.