As an antidote to the mediocrity and lack of clarity induced by corporate multitasking, Peter Thiel developed an alternate management approach when at PayPal: focus on a single contribution. This approach has a lot of benefits. It keeps people focused, takes advantage of convexity (accelerating returns) in difficult work, and removes excuses. It will also never work in the modern corporation.
“Why? Because Reality, that’s why.”
Here are three major problems with “one thing only” in a traditional corporate environment:
- There’s a lot of work that is important but of which there isn’t enough to make it a “one thing”.
- The question of who gets to set a person’s “one thing”– immediate management? upper management? the employee?– becomes extremely political.
- It eliminates the possibility for the employee to create redundancy in his “support network”, which is a necessity to keep the company stable.
Sporadic work
A lot of work is sporadic. It needs to be done, but it’s not enough to justify a full time, 2000-hour-per-year, effort. This is especially true in startups. The CEO can’t afford to neglect nasty paperwork because it’s not part of his “one thing”. Engineers shouldn’t take that attitude either. There’s a lot of sporadic work that’s extremely important, even if it fits poorly into a “one thing only” performance review process. For example, in a small company, recruiting is sporadic work, because there just isn’t enough activity to justify a full-time effort. Production support often becomes another category of such work, because most companies don’t want to have a dedicated person on it, because (a) it’s usually unpleasant, and people will expect very high compensation if doing it full time, and (b) that person can easily become a “single point of failure” for the business. When support duties rotate, they become another category of sporadic work. It’s not part of anyone’s “one thing” but it needs to be done.
Under a “one thing only” regime, sporadic work inflicts a performance penalty, because it means that less time can be dedicated to the assigned task. People who take it on tend to fall into two categories: (a) those who neglect it, because they won’t be rewarded for it, and (b) those who take it on eagerly and do it well, because they’ll be using that experience in a near-term job search. In other words, sporadic work is either done by people who are blowing it off, or people who are boosting their careers and likely looking for an external move– in part because a “one thing only” company makes it hard to move internally.
Who sets the “one thing”?
This is the most obvious problem with a “one thing only” management system. Who decides on a person’s “one thing”? Is it the immediate manager? If that, you now have a dictatorial company where a single manager dominates all of a person’s work. People are still willing to show up in the same place for nine hours per day, five times a week; but very few desirable employees are going to tolerate that sort of unilateral control over their time. Most employees are happy to follow orders from a manager in some of their working time, but sneak away 15 to 25 hours per week to support their careers through (a) side projects, (b) cross-hierarchical work that leads to internal mobility, and (c) online courses. The “fog of war” induced by corporate multitasking, incoherent direction, and convoluted command topologies is what enables them to get away with this self-support; remove that, and you have unhappy employees.
On the other hand, if the person chooses the “one thing”, then you have open allocation, whereby employees have the right to direct their own work. That’s great, but it’s inconsistent with a “one thing only” policy. Open allocation means that employees are trusted to change their “one thing” as needed, if (a) they believe a different “one thing” will add more value or be more likely to succeed, or (b) if they believe that they can effectively focus on two or more things and that it would be beneficial for them to do so. I like all of this, but it’s not consistent with a “one thing only” policy.
Under an open allocation policy where people’s creative energies are encouraged rather than stifled, one tends to observe an alternating pattern between divergent and convergent creativity. Divergent phases are characterized by searching and a process that appears disorganized. From a typical micromanager’s perspective, the person doesn’t seem to be producing anything, but is just having a bunch of conversations and reading books. This is an exploratory phase in which interest is often scattered as the person tries to figure out what to create. After finding something, a convergent phase sets in where the person shuts out all distractions and gets to work on that project. This is when the person is behind a closed office door for 12 hours per day, focused singularly on “one thing only”. Complex projects will usually involve several divergent and convergent phases in sequence.
Creative people know the often brutal process that’s required if one wants to achieve anything great. It starts with free expansion (divergence) and ends with brutal cutting (convergence). The problem with typical closed-allocation micromanagement is that it allows only for the latter, and the result is that workers end up “converging” on a product that no one (except the manager, who is often not a direct user of the work) really wants.
Redundancy and support networks
What keeps people in a company? Most HR theorists believe that it’s money, but there are a few problems with that explanation. No one has any real idea what technological or creative work is actually worth, but there is a fairly consistent market salary at any given time. Companies rarely depart from the market figure, the result being that most people can get a comparable job, if not a slight pay raise and a promotion, by changing employers. Money is a part of this, but what really keeps people working at the same place for years is differential social status (DSS): the difference in standing that exists between where a person is, socially, and where he or she would be in another company. This is strongly relevant because a person’s happiness while actually at work is much more strongly correlated to that person’s social status than compensation (which pertains to happiness outside of work).
A company owner has a great deal of DSS. He’s the boss, and if the company fails, he’ll be an employee somewhere else. People with a lot of connections and credibility within the company have high DSS. Most entry-level employees start with their DSS at zero. They probably have some social status (enough to justify having a job there) but no more than they would in another company. Incidentally, most employees find that, after a few years, their DSS at a given company is negative. They might have had a couple promotions, but their next big promotion is external. They leave.
What does this have to do with “one thing only”? Well, most employees in a company are going to put some effort into internal networking. They’ll take interest in cross-hierarchical work to make themselves visible to other teams. In part, this is about building up credibility, or DSS. More to the point, they also do this so that, if things go sour where they are, they can transfer. They’ll look for a “backup boss”– someone with managerial credibility, but outside of their direct reporting line, who can vouch for them– or five. In a traditional managed setting, the employee has serial dependencies: he depends on his relationship with his boss, who depends on the relationship with her boss, and so on, up to the company’s relationship with the market. Any one of these links failing can mean The End. It’s like ice climbing, in which you can do everything right, technically speaking, but still die if the ice calves because of an earthquake. Smart employees are going to build parallel support into their networks, building up a positive DSS and a web of connections to protect themselves if things get hairy.
In a “one thing only” company, this cross-hierarchical work is discouraged, at least for junior people. Their only allowed source of social status is the relationship with the manager. If the employee is tapped as a “protege”, then there is positive DSS, but managers rarely do that for obvious reasons (the other reports will call it “playing favorites”, which it is). Thus, the relationship with the manager does not provide positive DSS, unless the manager has unusually high clout. Typically, the manager/subordinate relationship delivers zero or negative DSS.
Solution
I don’t think a “one thing only” policy works, as much as I admire the endeavor to eliminate the mediocrity inherent in corporate multitasking. If I were running a company, I would adapt it as follows:
- Your job is to excel at (at least) one thing. You’re allowed to work on multiple things, but if you’re junior, you should have one main target (because excellence may be harder than you expect). You might also be mediocre at your assigned work, and that’s okay. The question I care about is not, “What do you do?” but “What do you do extremely well?” If you can’t answer that, the next question is “What will you do extremely well in the future, and what’s your plan to get there?” It might be something different from what you were hired for. That’s also fine. The first casualty of battle is the plan.
- Make your work useful. Finding something you excel at is important, but you also need to make your work useful to the company. If you excel at something that nobody wants or needs, then you’ll have to find something else.
- Excellence takes time, and that’s okay. No one’s expected to deliver meaningful weekly results. When you start, your job is to find something you can excel at, not show “contribution” from the first week, because everyone knows that real contributions start in the mid-single-digit months. Your task is to start looking. Talk to people. Figure out what you might be interested in, and what you might do well at, and go from there.
- If you try to excel and you fail, that’s okay. What you want to do may be impossible, or unaffordable. It may take five times as much time as you or anyone else thought. You may not have the ability or experience that’s necessary. Or you might produce something excellent that no one wants. If you do all this in good faith, you’re not fired. You shouldn’t even be embarrassed. You should be proud of having had the courage to try. Learn from the experience, write a couple internal papers to share what you learned, and move on.
- Impact reviews instead of performance reviews. You will never have a “performance” review. Why? Because the language of “performance” is fucking insulting. It means an employee isn’t getting the benefit of the doubt. Unless I know you extremely well (and if I’m running a company, I probably have too much on my mind) I am just not qualified to evaluate you for “performance”. Performance is personal and pejorative, impact is observable and objective. We will discuss impact. What is the actual effect of your work on the larger organism? The goal is to improve motivation (by reminding you that your work matters) and align interests, not to put you on trial. Also, whatever is discussed will never be used for personnel action (favorable or adverse). Its purpose is to remind you of the larger system in which you work.
- I will fire you only if you can’t excel. Low impact isn’t personal, and it’s not a problem if it happens in good faith. Most companies play “heads, I win; tails, you lose” by firing low-impact (or politically unsuccessful, which is more often the state of things) personnel while keeping over 99% of the proceeds (with the piddling “performance bonus” that is a tiny fraction of that work’s value) when their employees generate a high impact. That’s pretty goddamn scummy. Now, if I’m running a company, I am in the business of capturing the proceeds of creative, excellent work so I will be capturing and reinvesting the bulk of the profit. An employee who saves the company $10 million might get (in addition to a career-altering promotion) a $250,000 bonus– better than most companies give, but only 2.5%. However, if I’m taking the bulk of an employee’s upside in good, high-impact years, then I will also be taking on some of the downside in bad years. It’s only fair. So the only people I’ll fire are people for whom I cannot see any possibility that they excel within the context of the company. They will have three options. The first is a severance package, calculated based on the expected length of a job search, in exchange for non-litigation and non-disparagement. They can take that, resign that day, and leave in two weeks (to avoid the embarrassment of an abrupt termination). They will be allowed to represent themselves as employed during the severance period, and we will agree upon a positive reference, in order to ensure that they move along with their careers and have a chance to excel somewhere else (or not excel, as it’s no longer my problem). Their second option is to work out a transition plan, in which they hand over their duties properly, and which includes the allowance for time to conduct their job search on “company time”. I will even give them a private office where they can take phone interviews, and I won’t count on-site interviews against any vacation policy (if one exists). Their job is to find another job, peacefully and seamlessly. Their third option is to excel at something I didn’t think of, and prove me wrong.
The policy shouldn’t be “one thing only”. That can rapidly become a political mess. It should be: excel.
