useit.com Alertbox March 2004 Productivity

| Search

Jakob Nielsen's Alertbox, March 29, 2004:

Productivity in the Service Economy

Summary:
Yes, it is possible for white-collar workers to work smarter and become more productive. While intranet usability provides substantial initial gains, workflow usability can go much further and will save millions of jobs.

Not a day goes by without heated debate and laments about jobs being outsourced to offshore countries. Although my focus here is on the United States, these issues are equally relevant for other rich economies, such as those in Western Europe, Japan, and Australia.

Most commentators agree that it's preferable to solve labor market problems by increasing productivity, so that the value created per hour worked is big enough to justify the higher cost per hour worked.

Unfortunately, many commentators also think that while increasing productivity helped solve problems when we had a manufacturing economy, it's not going to work with the new service economy.

Example: Productivity Gains in Data Analysis

That said, the overwhelming majority of the effort required to produce a seminar lies in preparation, not delivery. I recently improved productivity for my current series of usability seminars by an estimated 17%, not by speaking faster, but by proactively accommodating changes in participants' registration behavior.

I did this by building a non-linear multiple regression model of the final registration numbers based on statistical analyses of the sign-up dates for 7,591 people who've attended my conferences over the last few years. This analysis estimated that our final audience would be twice as big as the room capacity in Melbourne, which was our only scheduled stop in Australia. To accommodate the increased demand, I added a series of seminars in Sydney.

Twenty years ago, I would have been forced to use horribly complex mainframe software to model this mid-sized dataset and would not have discovered the insights in time to book new rooms. Now, I use Excel's explorative data visualization features, coupled with custom-built formulas in its primitive spreadsheet language. Having the data at my fingertips in a visual PC interface let me torture it until it confessed.

Information Technology and Productivity

For intranets, we know that good design can double employee productivity. This estimate comes from our intranet usability testing, where people using the worst 25% of intranets required 99 hours per year to perform typical employee tasks, whereas people using the best 25% of intranets accomplished the same tasks in 51 hours per year.

Of course, even if all companies improved their intranet usability, the resulting productivity gains would only apply to the time spent using the intranet. On the other hand, this doubled productivity estimate is nowhere near intranet usability's ultimate potential. On the very best designs we tested, employees could complete their tasks in 27 hours per year, and there is no reason to believe that today's best designs are the last word in intranet usability. We're likely to see even better designs in the future, leading to even greater productivity gains.

It's great to employ usability methods to simplify intranet user interfaces and expedite performance within existing tasks. Do it. Millions of dollars are waiting to be saved for the average company.

But, as the saying goes, we should also work smarter. Enterprise software has largely failed until now because it's cumbersome and it automates awkward and inefficient procedures. Business process reengineering must become more than a slogan. In particular, we must change how we define it; rather than simply "automating existing processes" we should be "designing workflow that's optimized through computer support."

Through methods such as field studies, task analysis, and user testing, organizations can discover new ways of working and better ways of supporting work with information technology. In particular, immense gains are possible through better collaboration interfaces, better knowledge management, and better decision support. Most existing systems are unworthy of their names: they don't help people collaborate better, they don't increase knowledge utilization, and they don't support better or faster decisions. But they could. That's the next frontier for enterprise usability.

Growth in Usability Jobs

One positive aspect of outsourcing some programming jobs is that software development will get cheaper. Applied properly, the savings could improve usability and thus the productivity of end users. One of the main reasons for poor design is that design teams lack the resources required to implement all of the usability recommendations. (The biggest reason for this is that many companies do zero usability, but even companies committed to usability never get around to implementing every guideline.)

Ultimately, great information technology requires much more than 10% usability. To exceed that, we'll likely need as many people working on user research as there are people working to implement usability findings. So, in the long term, we might need two million usability workers in the U.S. alone. They won't all use the same methods as current usability professionals, and many of them may have different job titles. But to build enterprise solutions that are good enough to truly boost American productivity, we'll surely need to create many more usability jobs than we lose in outsourced programming jobs.

Usability is key to increasing the service economy's productivity, because only attention to the way humans work can help them work smarter. If we adjust our focus accordingly, we won't just save billions of dollars from productivity gains -- we'll also save millions of jobs and create millions of new ones.