Skip to main content

A Rebuttal to Michael Lopp's "The Engineer, the Designer, and the Dictator"

Posted in

I just got back from DrupalCon Portland. I loved it, as I've loved every DrupalCon that I've had the good fortune of attending.

This year, as with every year, there were a couple of keynotes by big-name speakers outside of the Drupal community. This year, the speakers were Karen McGrane, who was great, and Michael Lopp (aka Rands) who was... definitely somebody who spoke at this conference.

From what I saw on Twitter and heard throughout the conference, I'm not the only Drupalista who was less than thrilled. But I also heard some attendees express that they felt that he had good ideas, and that we as a community should pay attention to them.  To that I say: What good ideas?

Briefly, Lopp's thesis states that there needs to be somebody in three different roles in order for any project to be a success, and these three roles are the titular engineer, designer, and dictator. The engineer brings order and solves problems, the designer humanizes the project, and the dictator provides vision and keeps the project on track.

The talk itself contains a lot of misleading rhetorical tricks and sloppy stereotyping, but I want to focus on what bothered me the most: the third role, the Dictator. The Dictator is the person who imparts vision to a project, gives it coherent direction, and basically makes sure that all parts of the project stay focused on basic goals. It's not hard to draw a line between this fictionalized role and the role of the CEO in the business world. Lopp himself does this. Bill Gates, he says, was a dictator. Steve Jobs was a dictator. I can follow Lopp this far. Those two were extremely visible CEOs of companies that came out with revolutionary products.

But Lopp also provides a bunch of counterexamples. He comes up with a bunch of companies and CEOs which he claims don't have dictators. Following Lopp's own logic, these are the companies that don't have successful products, because they don't have dictators.

Steve Ballmer's Microsoft is first up.  Since Bill Gates left Microsoft, everybody's favorite evil empire has gotten significantly less sexy, but they continue to put out solid numbers and good products.   Windows 7 was a critical and sales success, and Windows 8 has even jumpstarted a graphic design trend. Even Lopp himself allows that the Xbox was a good product, though he handwaves that away by saying that there must be a Dictator somewhere else in Microsoft which sounds an awful lot like affirming the consequent to me.

Oracle, IBM, and HP are next for non-dictator companies. According to Lopp, these companies' success (or non-success, in the case of HP's fallen valuation) is due to them using business models, not creating products.  I'm not sure what Lopp thinks that IBM and Oracle's enterprise offerings are, but they are most assuredly products, and they are definitely successful.

But the absolute most perplexing examples that Lopp uses are Facebook and Google. According to Lopp, Sergei Brin and Mark Zuckerberg are not dictators -- they're engineers masquerading as dictators. These are the companies behind some of the most successful products of the last 20 years. Facebook's platform drives the social web and has enormous success, and Google's mail, mapping, and mobile devices are assuredly successes. 

And finally, the category that Lopp never brings up (until the Q&A, when he quotes some factually inaccurate figures about Drupal) is the open source model. The various distros of Linux have quietly eaten the server world, and open source languages, frameworks, and databases like php, Ruby/Rails, Wordpress, Drupal, and MySQL run most of the web. None of these companies have anything that can be qualified as a Dictator. The open source model is, in fact, inherently opposed to the idea.  And yet it has proven to be an enormous success.

I understand the conception of Dictator, and I understand why Lopp separates these companies into Dictator haves and have-nots. But how on earth does this arbitrary distinction prove any kind of a point about product success? How does he point at all kinds of successful companies and insist that they don't have successful products?

This brings me back to my original point: What good ideas, exactly, did this keynote have?

The best argument in favor of Lopp's keynote is that the roles he describes aren't meant to necessarily map to a single person, but that there exists some kind of generalization about a project which both allows one person to fulfill several roles and allows one role to be filled by multiple people at the same time.  But, then why does Lopp only use single-person dictators in his examples?  And what's the point of the model, then? Is the model really saying nothing more than that every successful product has a cohesive direction? Isn't that's embedded in the definition of successful already without us having to be spoon-fed that at a keynote?

In the end, it was a keynote that used weak, anecdotal evidence to reinforce lazy oversimplifications, and then overgeneralized those ideas beyond any kind of useful application.  This is really just masturbatory executive self-justification of the Big Man theory.