How technical should executive management be?

You’ll find a wide range of technical depth in executives who oversee software engineering, both at the VP level and the CEO level. I find one CEO transition illuminating of one end of that range: the replacement of Eric Schmidt by Larry Page as CEO of Google.

We think of Eric Schmidt as the polished adult supervision, and people said it was time to return to leadership by a founder who was closer to the technology. So who is this “adult supervision”? Well, he has a PhD in Electrical Engineering and Computer Science from Berkeley, and if you’ve ever written a compiler, he probably made your job a little bit easier by creating lex (see his description here).

On the other hand, I’ve worked with VP’s of Engineering who I’m pretty sure couldn’t tell C code from HTML. (It did not work out very well).

What’s the right answer? History would show that CEOs who understand technology are radically more effective (see Bill Gates, Larry Ellison, Steve Jobs, Eric Schmidt, etc). I generally lean strongly towards highly technical management. Part of my argument is historical, but I also have a first principles reason: identifying talent (and non-talent).

My rule of thumb:

  • A VP of Engineering can be a little rusty but ought to be, underneath the rust, pretty good (but not the best) at the most technical work that goes on in the team. Why? He needs to be able to identify talent; in practice this means discriminating between a very good and an average engineer in any team.
  • A CEO spans a number of areas and has VPs in each who should be able to spot talent. But he should be able to identify problem employees himself to keep his VPs honest; in engineering, he should be able to discriminate between an average engineer and a very bad engineer. (Though this is a discussion about technical knowledge, in my opinion he should also be able to discriminate between an average and a very bad sales rep, finance person, and marketing person.)
It takes senior leadership to hunt down mediocrity and to retain talent. Fully crowdsourcing your talent management by following reputations alone may be your best strategy if you don’t understand the work your team does well enough to make up your own mind, but its a poor substitute for having an opinion (which can also be informed in part by reputation). I remember getting a phone call from a very talented engineer who we will call Mark (in this case working in a consulting organization) who was shocked to learn he’d been laid off. In my opinion Mark was easily in the top 5% in his organization. Why was he laid off? Because on the day his VP, who we will call Gary, needed to cut 10% of his team, Mark wasn’t billable on a customer engagement and Mark was highly paid. I suppose Gary didn’t bother to think about why the managers in his organization thought they ought to pay Mark so highly before Gary took the opportunity to cut some “fat” from his organization. Little did Gary know he was cutting brain. (Well, the brain is around 60% fat…)
You’ll have trouble holding on to the talent that your executives don’t recognize, and you’ll have trouble eliminating the dead wood that your very top executive doesn’t recognize. So I think you should have executives who know the difference.
I recognize that my CEO description bounces a lot of very good CEOs. John Chambers at Cisco would be one notable example. So like every good rule, there are exceptions 🙂

— Max

6 comments so far

  1. Cliff Elam on

    I think your examples of technically led companies are very good, but I suspect that a good methodical drag across companies might turn up a different distribution of success versus background.

    My experience with technical management is much different, I guess. My most memorable managers have been non-technical. Or at least the areas where they have impressed and influenced me the most have been from the non-technical portion of their experience/knowledge. (One was, and I am not joking, an ex-plumber and karate trainer who ran an engineering group. Oh, and he was Polish. Life is peculiar.)


    • Max Schireson on

      Yes, there are some counterexamples and my methodology is far from scientific. Lou Gerstner did a great job at IBM, for example.

      One of the things which makes it more complicated to assess precisely is that many of the technical CEOs (Eric Schmidt being a notable exception) are founders. Its possible that what’s great about them is not their tech background but other founder-related attributes, eg as Ben Horowitz discusses their long term view, moral authority, and complete knowledge of the business.

      Certainly technical knowledge is not enough to be a good manager. I can see another blog post there…

      — Max

      • Cliff Elam on

        It would be a good post, and I’d like to read it.

        I’ve always felt that the attributes of a really good technical leader were somewhat at cross-purposes with the necessary leadership qualities for everything else. For me technology was always about the drive for certainty – this algy is the right one to use in these X situations. But getting people into high performance mode is all about living with a lot of grey ambiguity. Even seeking it.

        Luckily people can believe two impossible things before breakfast….


  2. Subraya Mallya on

    Hi Max
    I could not agree with you more. For a technology company, it is imperative that the CEO be technical and better yet own the Product Vision. Adult supervision can be had with a great board of advisors who are un-biased and have your good interests in mind. In addition to identifying good from bad in engineering, I also think to a large extent attracting/hiring technical talent is done based on leadership. Technologists usually like working for someone who is herself a technologist. Meshing the Product Roadmap with technology shifts is something a non-technical CEO cannot think and it leads to lot of friction. To your point of Lou Gerstner, I think IBM is a odd-ball where each GM of a business unit performs the role of a CEO so it made sense to have a head coach in Lou Gerstner. Also during his tenure, they focussed more on services and catalog sales of hardware. For software companies it is an imperative that a tech CEO lead the charge.

    BTW, big fan of MongoDB and in the process of moving from PostgreSQL to Mongo.

  3. Dwight Merriman on

    is the context here pure tech companies only?

    • Max Schireson on

      Mostly. That’s what I was thinking about. At some point I’ll add some thoughts about other areas.

Leave a Reply

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

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

Facebook photo

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

Connecting to %s

%d bloggers like this: