Friday, November 18, 2011

Brainpower isn't scalable

This pretty well describes my job, although, I don't work for EDS.

However, I'm always running up against three limits: I lack the ability to forecast everything, the ability to know everything currently going on, and the ability to fix everything that has gone wrong. I use a lot of data tracking tools, some given by HQ, some inherited from others built for their needs, and some built from scratch for myself. I share a lot of data with my coworkers, and am working constantly to refine what data is useful to them immediately, and what data is useful for them to forecast their needs. I track a lot of data to try to build prediction models based on past history.

Here's the hardest part, the source of all my problems: math is merely a language we use to precisely describe the universe around us. It represents many precise things, some accurately - but it is not reality, merely the language we use to describe reality. For each number I deal with, there is a customer somewhere who wants something. And customers are people. For each number in the system, there are employees working to move product to that customer. Employees are people, too. I speak the language of math all day, in every tool, but the moment I forget that the entire systems is built by people, run by other people, for the benefit of yet more people, all the mathematical genius in the world will not save me. Every number I change represents people being told to do something - and there is no formula in the system that can turn an equation red to warn me I'm going to frustrate people, and frustrated people work more slowly. There is no tickybox I can check to improve morale, but improved morale will definitely show in more things happening faster. Mentoring shows as a slowdown in efficiency - a highly efficient person is doing less of their own work - but that slowdown pays off by greatly reducing error rates forever after down the line with the person mentored.

Now, I'm out with my people. I'm running - literally - around trying to account for all sources of critical or urgent (or critical and urgent) errors, and make sure that as the work scales, the errors will not scale correspondingly (or logarithmically). I'm answering questions from the newest employee trying to figure out his break time to the project manager wanting to know status on major department goals, while trying to actually do my job description. The more people we hire, though, the harder it gets to know all my people, and know all the problems, and anticipate what people will do.

But sooner or later, I'm going to hit the limits of my brainpower. You see, I'm human. I can only hold so many things in my mind at one time. I won't get smarter as the problems get bigger. I can delegate parts of problems, but delegation means training and trusting that a subordinate will know how to handle everything they're meant to, and when they're not meant to, and how to deal with the things they're not supposed to handle or fix themselves, and the consequences of taking initiative. That only works so well... I mean, Jesus chose twelve disciples, and one of them was a saboteur, one denied any affiliation, and ten ran from the pressure. If God incarnate in flesh couldn't find and train 12 people to do what he wanted, there's no hope I'll ever have a perfect team. The Catholic church claims direct apostolic inheritance, and we see how badly they've screwed up over the years. No matter how hard I try, sooner or later, I can't control it all, and if I try, I won't be able to respond fast enough... and no matter how hard I try, the customers are going to do whatever the customers want to do when they want to do it, and no plan can simultaneously keep us at top efficiency for lots of customer demand and little customer demand.

There's only one solution - and it's the long-term, time-intensive one. Train people to do the best job they can, educate them on the common errors and how to fix them, and then give them incentives to work as hard as they can, while getting out of the way and letting them thrive. No matter how seductive the statistics, no matter how simple the mathematical solutions seem to be, it's not numbers or turtles all the way down, it's people... and that's why we have to put our money on Adam Smith's invisible hand(1), on free-market and free will, not command and control.

(1)"But it is only for the sake of profit that any man employs a capital in the support of industry.... As every individual, therefore, endeavours as much as he can... to employ his capital... that its produce may be of the greatest value; every individual necessarily labours to render the annual revenue of the society as great as he can. He generally, indeed, neither intends to promote the public interest, nor knows how much he is promoting it... By directing... industry in such a manner as its produce may be of the greatest value, he intends only his own gain, and he is in this as in many other cases, led by an invisible hand to promote an end which was no part of his intentions. Nor is it always the worse for the society that it was no part of it. By pursuing his own interest he frequently promotes that of the society more effectually than when he really intends to promote it."


  1. That may be the greatest description of the organizational growth vs experience vs delegation problem that I've ever seen. Ever.

    And every bit of it is earned; I can tell that, too.

    Well done, ma'am.

  2. I would hazard a guess, based on your post, that you're trying to supervise too many people.

  3. Rev Paul - thanks. I'm not sure it's terribly coherent, but I'm glad that you understand.

    Comrade Misfit - I'm hazarding several guesses, and trying to gather the data to test the hypotheses. One is that I'm supervising too many people. Another is that I'm supervising too many new people, with a highly skewed ratio of new to experienced. But if I have too many new people, is it that they're new, or that we have insufficient training and mentoring (and thus a lower efficiency and higher error rate)? Is it that I'm failing to delegate too many things? Or will the problems I'm currently facing scale with the work, which means that I will be working as hard or harder even when I have an assistant on my current problems, and have the problems inherent in delegation as well?

    Or is it that I'm not using all the tools available to the full extent, and thereby making my life harder? Or that I'm just not that good at my job?

    All I can do at present is keep my department going, track the numbers, and based on human input, feedback, gut feel, and the numbers, constantly try to improve things for the long term. It's people all the way down, though - math can tell me what happened, and where, but it can't tell me why, or how to fix it.