Via Todd Gailun and former HP CTO Phil McKinney, I came across Leslie Perlow's short piece on how to "Overcome Your Work Addiction" today. 

Consider the following: Works long hours. Carries wireless device everywhere. On the phone at kid's soccer game. Checks in frequently over vacation.

Does this describe your life? If you're like the hundreds of executive education students I teach each year at the Harvard Business School, you point to the hours you work, the places from which you work (even on vacation), the times at which you work (even when supposed to be spending time with family and friends), the fact that your wireless device is never far from reach, and declare without any hesitation that you're always "on". And, you probably declare yourself an addict.

But what are you addicted to? Your wireless device? Work? These are the most common suspects, but I would argue that many — if not most — of us are addicted to success. We are successaholics not workaholics. We're obsessed with work because of the satisfaction we get from the kudos for achievement, not because of some deep-seeded satisfaction from working long hours, as an end in itself.

Perlow goes on to describe a project she did with Boston Consulting Group in which people regularly went offline, which has become the basis for her new book, Sleeping With Your Smartphone.

I certainly am in favor of anything that gets people to think more deeply about how they work, where the satisfaction (or dissatisfaction) with their work comes from, and how they can improve their lives. But I wonder if this diagnosis is a bit too flattering: You're not addicted to technology, Harvard Business School executive ed students (that would be unflattering)! You're addicted to success! You like achievement too much!

This is easy to parody: one friend summarized the argument as, You're not the tech industry's bitch, you just don't know when to stop being awesome, which maybe is taking things a bit far. (Though one commenter's point that this might not be, but "'fear' of being dispensable" is also a good one.)

But I think there are a couple valuable things embedded in Perlow's study that I think are worth drawing out.

First, it seems to me that people aren't addicted to success, but to the feeling of success. There is an important difference.

In a recent study on multitasking (thanks to Maria Panagiotidi for the pointer), Ohio State University researchers Zheng Wang and John Tchernev studied college students' multitasking, and found that "multitasking often gave the students an emotional boost, even when it hurt their cognitive functions, such as studying." As they explain in a forthcoming article in the Journal of Communication (aren't all journals communicating?),

The increasing popularity of media multitasking is frequently reported in national surveys while laboratory research consistently confirms that multitasking impairs task performance. This study explores this apparent contradiction. Using dynamic panel analysis of time series data collected from college students across 4 weeks, this study examines dynamic reciprocal impacts of media multitasking, needs (emotional, cognitive, social, and habitual), and corresponding gratifications. Consistent with the laboratory research, cognitive needs are not satisfied by media multitasking even though they drive media multitasking in the first place. Instead, emotional gratifications are obtained despite not being actively sought. This helps explain why people increasingly multitask at the cost of cognitive needs. Importantly, this study provides evidence of the dynamic persistence of media multitasking behavior.

In other words, multitasking is emotionally appealing even when it's functionally destructive. As they explain:

Then why do people increasingly multitask at high cognitive costs? This study suggests at least two reasons. First, although cognitive needs are not gratified by media multitasking, emotional needs are, such as feeling entertained or relaxed. To add a twist, emotional needs are not actively sought in media multitasking. Thus, emotional gratifications appear to be a “byproduct” obtained from the behavior. How does this occur? While our model cannot provide information about specific activities, it suggests that if participants were, for example, studying for a test while watching TV, their multitasking might lead them to feel satisfied not because they were effective at studying, but rather because the addition of TV made the studying entertaining. In the long run, it is likely that this emotional gratification associated with multitasking serves as an implicit yet powerful drive, similar to the formation of implicit attitudes through classical conditioning (Olson & Fazio, 2001), to engage the students in media multitasking again and again. In this sense, the “myth” of multitasking actually is partially caused by the “misperception” of the efficiency of multitasking and by positive feelings associated with the behavior, which is emotionally satisfying but cognitively unproductive. Meanwhile, it is interesting to note that emotional gratification itself is a function of emotional needs and multitasking (Figure 3). Multitasking increases emotional gratification when emotional needs are low, but it decreases emotional gratification if the needs are high.

Second, habits play an important role in media multitasking behavior, and multitasking can be self-reinforcing. Our findings show that habitual needs increase media multitasking and also are gratified by multitasking. More importantly, our dynamic analysis found a significant feedback effect of the media multitasking behavior. This feedback effect integrates past media multitasking experience into the current situation, and accumulates all the exogenous influences from needs and gratifications to reinforce the behavior. In addition, needs and gratifications themselves are self-generating and self-reinforcing as indicated by their own significant feedback terms. Thus, the habitual continuation of media multitasking behavior is further strengthened, in a more complicated way, by self-reinforcing needs and gratifications.

I would theorize that there's a similar thing going on with Perlow's group (or maybe the following only holds true with always-on workers who are mere mortals, rather than consulting world demigods).

In services and the delirious professions, it's easy for being always-on to be an end in itself, proof to yourself and others that you're poised at the starting line, ready to spring into the client's service at a moment's notice. It's easy for this kind of accessibility becomes a product of its own, divorced from any kind of real value: I have lawyer friends who are certain that their instant accessibility to clients is not only unnecessary but makes it harder for them to serve their clients well, but the reality is that clients like the idea that their high-priced counsel is available at a moment's notice. The client feels like those fees are being earned, and it makes them feel important.

Now, of course you have to be available when things blow up. But it's easy for the accessibility to become an end in itself, and to get in the way of doing more substantive things. I think that's what happened to me when I was at my most multitasking-intensive worst, when I felt like I was starting to lose my ability to think really serious thoughts.

Being available should be a condition, a stance; it shouldn't be something that requires you to be checking your cellphone at kids' events. The fact that it's something we have to do, makes it easier to confuse with actual working.

It's also worth noting that there's no clear limit to always-on-ness. How do you know you've been always on enough? For most of us, there's always more work to do, more calls to return, other things we could start. The extrinsic measures all point in the same direction: Do More.

This is where the second insight from Perlow's work comes into play. People in BCG went offline when their colleagues did– when the social norm was formally changed so it was okay to be offline at certain times:

Starting at 6 pm, for that one night, they were to do absolutely no work — not even to check their wireless devices. They were to completely disengage from work. Each person's night off was set well in advance and was not supposed to change, even if suddenly there was a client deliverable the next day. And, each week the team met to discuss their progress, with each team member being required to share whether they had taken their night off, and if not, why not. Suddenly, always being on was not the badge of honor that it once had been. Rather, team members were publicly applauded for taking their time off — even the night before a major deliverable — and they were shunned for failing to take their nights off.

This highlights the degree to which things like the Digital Sabbath require either a very high degree of intrinsic motivation (something that the people I interviewed all had), or a cultural shift in your core group– the people who are most likely to notice if you're connected, and to take advantage of that.