A world Without Email: Reimagining Work in an Age of Communication Overload

The Specialization Principle
A Productivity Puzzle
In his wide-ranging 1996 book, Why Things Bite Back: Technology
and the Revenge of Unintended Consequences, the independent
scholar and writer Edward Tenner tackles a “productivity puzzle”
that’s proven both significant and widely overlooked: Why did the
arrival of personal computers in the workplace fail to make us as
productive as we predicted? As Tenner writes, “the huge investment
in computing in the 1980s and early 1990s” made office workers feel
“autonomous, in control, more powerful, and absolutely more
productive.” It was compared to a second industrial revolution,
something that would transform work in profoundly positive ways.
“But toward the end of the 1980s, the sentiment grew that something
was not right,” and by the early 1990s, people “from within
technocratic culture”—economists, business professors, consultants
—began to notice that the computer’s predicted gains were not fully
materializing.
1
This skepticism was sparked in part by discouraging data.
Tenner cites a study by the economist Stephen Roach which found
that between 1980 and 1989, investment in advanced technology in
the service sector grew by over 116 percent per worker, while the
workers’ output increased less than 2.2 percent during the same
period. He also cites a study by economists at the Brookings
Institution and the Federal Reserve that calculated the “contribution
of computers and peripherals as no more than 0.2 percent of real
growth in business output between 1987 and 1993.”
2
Even without this data, many people were coming to similar
conclusions about the failed promise of the PCs that had become

---

ubiquitous seemingly overnight. These shortcomings were
particularly noticeable in professions that existed both before and
after the computer revolution. My grandfather, like me, was a college
professor. I spend most of my day interacting with a powerful
portable computer equipped with high-speed wireless access to the
internet. My grandfather, by contrast, didn’t buy his first computer
until after he retired (I helped him set it up), and there was no
evidence that he ever actually used it. He would transcribe his books
on yellow legal pads to be later typed up by an assistant, and he
didn’t need the internet for research purposes, as he packed his office
with a massive personal library covering the topics he studied. I can
highlight many small-scale tasks in my life that my computer has
simplified. But if we look at the big picture metric that matters most
for scholars—research output and scholarly impact—I can’t argue
that I’m more productive than him, especially considering that he
wrote many books and ended up with an endowed chair in religious
studies at Rice University before ending his career in the role of
provost at a large theological seminary.
Tenner offers several explanations for this puzzle, but one of his
primary arguments is that instead of reducing labor, computers end
up creating more work. Some of this extra work is direct. Computer
systems are complicated and change every few years as existing
technology becomes obsolete. They also break a lot. The result is a
large time investment in learning new systems and trying to get them
to work. Around the time I was writing this chapter, for example, my
speaking agent stopped by my office for a visit. As we were talking
about workplace inefficiency, he told me the story of the woes the
agency was facing trying to get a customer relationship management
system called Salesforce working properly for their particular needs.
After endless hours of tinkering, they ended up having to hire an
additional expert to do nothing but work with the system. My agent
wasn’t so sure they were really gaining a productivity bump
compared with the old days of Rolodexes and business cards.
More insidious, however, are the indirect labor increases
personal computing instigated. A major problem with personal
computers, Tenner notes, is not that they make individual tasks
harder, but that they make them just easy enough. To explain, he
points to a remarkable 1992 paper, published in the journal National