Blogs

Well it’s been about a month now since I attended WebVisions 2005. I attended an interesting talk by Chris Bond, a usability analyst at Portland General Electric. The topic was how usability improvements saved PGE approximately 250K or the equivalent of 10 FTEs (full time employees).

The presentation started with a video clip of the users of new a new client-server win app that was used in the call center. Most of the users expressed extreme frustrations, one user called it a “time incinerator” and “this sucks” was the comment from yet another dissatisfied user.

Usability is the key to customer satisfaction, you wouldn’t live in a house if the lights, plumbing and door wasn’t there…I know!?

I jotted down some points from Chris’ presentation, my thoughts included in some of the points.

Classic usability problems
Takes too much physical and cognitive effort
Managing multiple windows on the desktop
Switching between keyboard and mouse, this takes time and is annoying (can also cause carpal tunnel syndrome)
Switching between typing and hand writing notes (the system should data so users don’t need to do this)
Superfluous keystrokes, unnecessary dialogue boxes (i.e too many “ok” pop ups)
Unnecessary task complexity
Unnecessary data re-entry, “swivel chair” integration (manually visually re-entering data from multiple screens, too much copy and paste, etc..)
Compartmentalization of data – transaction based
Inefficient task performance (multiple ways to do same thing – users often choose the least efficient somehow)
Pre-populated data lacking. such as: default date, employee/user number, customer/invoice number…
Default option is cancel in dialogues

Some tips on how to build a usability process: Look at baseline metrics, how long to complete tasks
Sit with the user and visually observe them
Develop top ten list of usability problems
You need a robust fully functional prototype
Solve come up with solutions to the usability problems and re-test, repeat
Measure number of steps
Measure times of tasks – before and then one month after implementation and trained users
Begin timing when tasks start
Suspend timing for questions
End timing when task is completely done
Throw out aberrations
Use same users from baseline to improvement (studies)
Note issues
Focus on routine tasks

Some GUI design tips for improving usability: Navigation task buttons
Pre populating fields
Auto tabbing
Automated calendar
Summary screens – user needs don’t map to data model and hierarchy “summary screen example”
Removing the need for a mouse for high powered users a good thing (doing this on the web is challenging though)
Insulate user from destructive actions – forced, background save

Usability / HCI metrics:

keystroke on standard board 0.28 seconds
move hand to mouse 0.36 seconds
move cursor to new location on screen 1.5 seconds

Business benefits PGE got from the usability analysis:
shorter hold times
decreased abandon rates
increased accessibility
improved service levels
shorter talk time
Savings of 170K = 5.5 FTE (Full-time employees) *I’m not sure why this number is different that the 250K why an employee is 30K per year, I’m guessing that’s a standard in ROI calculations.

Observations about users even with a more usable system:
Old behaviors die hard (training and new system isn’t enough)
Monitoring and coaching is required
Practices need closer attention

In conclusion Chris noted there is often hesitance to usability analysis and improvement for various reasons. One is the original developers/designers/managers are afraid they will be made to look bad. This would be a great argument for doing the analysis up-front and having a usability focus from project inception.

One opinion that I don’t agree with Chris on is that usability issues are defects and is not something you should have to pay for. Coming from consulting background, everything should be billable, how is this made objective, usability seems subjective on some levels. Even if it’s not, if the customer doesn’t want time spent on it in the project budget then they should if they want it retroactively. It takes a long longer to make something complex like a software system simple and easy to use. BJ Fogg and Phd from Stanford said it takes 5 times as long to make something simple.

It’s not always easy to charge for and sell usability, it should a separate deliverable in your statement of work – it’s quicker and cheaper to just slam a system with a database and a bunch of screens together without thinking about usability and work flow, and clients always demand quicker and cheaper! Looking for some ideas on this…

It was also mentioned that remote usability testing using software installed on end users computer tracks mouse clicks and screen captures is not as valuable as in person observation and interviews.