On productivity

Productivity is almost a modern-day version of virtue. It is quite common to look up to productive individuals and feel bad for not being productive enough. I certainly value productivity myself, but I think it is easy to get confused about what it actually means, which is another property it shares with virtue.

I spent 2011-2012 working at Google Research. At Google, it is obvious what productivity means: the company has clear goals, employees have clear tasks on projects which are well-defined. That is true even in Research. Being productive means advancing the company's goals and your projects at a high rate. Usually, working long hours leads to getting much useful work done and therefore directly to productivity.

Since this relationship between time spent working and productivity is often true in industry, the two have become almost synonymous. In this post I'd like to explore how they are not always synonymous, and possible implications.

I am not talking about the value of taking breaks and not working all the time. That taking breaks can increase the total amount of work done is obvious and frequently discussed. I am talking about something else altogether.

I am now back at Stanford, working on finishing my PhD. At Stanford, and indeed in academia at large, the same identity between hours of work and productivity is usually assumed. And the strong correlation between time spent working and work generated generally still holds. But in academia, another factor becomes important, a factor which separates work generated from real productivity. I refer to this factor as the energy conversion efficiency of labor, or EEL in short.

What EEL means is that not all the work you generate is actually valuable. In other words, you can produce a lot without being productive at all. And this is much more pertinent to academia than it is to industry.

Let us define EEL intuitively as the fraction of work produced which is actually valuable and useful. Then I formulate the following 80/20 "rule", because it is fun to make up 80/20 rules and it saves you the trouble of estimating precise quantities:

In academia, over 80% of the work done accounts for less than 20% of the value generated. In industry, 80% of the work done accounts for 80% of the value generated, and the remaining 20% is probably employee 20% time, which is often valuable in less obvious ways.

So why is so much real work done in academia wasted? There is no single answer, but many factors contribute to this low EEL. I will only touch on one factor in this post, and leave the rest for further discussion.

The publication system so ubiquitously wastes practically everyone's time that it has to be the one I use as an example here. Let us assume that research itself is completely productive, and examine the process of publishing it. Writing the paper is productive if that paper gets published and people read it. But papers are often rejected, usually after many months in review, during which the authors usually forget all the small technical details of their work (what each line of code does, etc.). Addressing reviewer comments is often more time-consuming and arbitrary than its actual value in making the paper better. Resubmitting to a different journal involves much restructuring, rewriting, changing file formats, writing new cover letters, and similar elements of real work which are nevertheless not generating new value. In the worst cases the work never gets published at all, in which case if it had any value at all it is bound to be repeated in the future, which makes it wasted work. I will not say more about this since I believe that everyone agrees that the amount of work done on getting research published is much greater than it should be.

The final point I'd like to make here is about what this implies. I spent much effort, as many others, trying to sustain long hours of work and get more done. I even built a web application (Quantified Mind) to help people find out how to improve their cognitive performance, which is usually with the goal of getting more work done. Sustaining 15-hour workdays and so on. And this makes sense when EEL is high and most of the work you do produces real value.

However, when EEL is low, the whole game needs to be reconsidered. If you work 15 hours a day at 20% EEL, you are less productive than working four hours at 80% EEL (remember, I am talking about actual time spent working, not just time "at work"). This is probably why so many successful people don't actually work that much, but still produce much value. I believe the right strategy is to take time away from execution to planning: before doing 15-hour days, spend much more time carefully choosing what is the work you are going to spend time doing. Perhaps do a bit less, but make more of it matter.

What do you think, dear reader? Am I right or am I missing something important? What other factors reduce EEL in your field? What is the EEL of this post?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s