Corvus Consulting

I came across a short post on a Google-watching blog that draws from some statistics on the most commonly-used features in MS Office:

Jensen Harris, Group Program Manager of the Microsoft Office User Experience Team, published in 2006 a list of the most used features in Microsoft Word 2003, according to data collected from the users who opted for the Customer Experience Improvement Program:
1. Paste (11% of the usage)
2. Save (5.5% of the usage)
3. Copy
4. Undo
5. Bold

These five commands account for 32% of all the command usage in Microsoft Word 2003, as they are used very often.

The post goes on to say that as a development strategy, Google’s office-apps development teams could use those stats to refine the few operations that are used most, rather than trying to cram in more numerous and more complicated features. On the surface, this makes some good sense: follow what people are doing most with your apps, and support them as much as you can. But it misses some of the deeper thinking that can make the difference between an application that does just enough to one that truly evolves.

One thing that relying too much on statistical usage data can get you is a blind spot for parts of the application that are important, despite a low frequency of use. An anonymous commenter points this out with a gentle poke:

This is like saying that most time spent in car involves going and only a little stopping, therefore stopping is not important!

True, that. Sometimes it’s useful to take an idea to an extreme to show its weak spots. The more problematic thing with making plans by stats alone is that they opt for incremental gains in efficiency or raw power without stopping to ask the question: why do people use this feature so much?

Probing the reason for high frequency of use can reveal opportunities to solve underlying problems. That pasting is the #1 feature suggests that data isn’t moving from one document into another, or replicating through the same document, intelligently enough. A user who repeats a Paste command over and over might not be hip to the magic of Find and Replace. That Undo is at #4 suggests that people are making a lot of mistakes or changing their minds. It could be that the natural number of typos pushes Undo to the top five, but it could also be that usability issues are leading to mistakes or unwanted results. Making it faster or easier to Undo a mistake isn’t better than preventing the need to reach for Undo so often.

To appreciate this, consider how Google Docs does solve an underlying problem of trust in the software to get around the need to click Save so often. Again, from the post that kicked this off:

Google Docs auto-saves documents so you don’t need to press the Save button…

Provided that immediate saving doesn’t create other problems, like exposing thoughts before they’re ready to show, this is a perfect example of solving the problem rather than just making it easier to take the action.

Stats can be very useful, but they’re misleading without an analysis that questions the reason for a trend. Spending some time to ask why a feature is needed so much in the first place can reveal the way to a better experience, not through statistical gains in efficiency, but by removing the need for the feature altogether or showing a better way to the same result. After all, the best solution doesn’t make problems easier to solve, it makes them go away altogether.