Adam Lashinsky wrote a must-read article on this week's Fortune (Google's biggest threat? Itself, 5/26) for students of innovation. A must-read because it provides a timely and necessary antidote to the causality-challenged study and writing about innovation. 

Since Google's IPO–the requisite proof positive of Google's genius–there have been countless articles dissecting the search companies innovative techniques for managing innovation (see for example, BW's How Google Fuels Its Idea Factory, 4/29).  Everything from the small groups free to develop (and launch) new applications to the free haute-cuisine to the 20% time programmers are expected to devote to their own pet projects. To anyone who has been around the valley for over a decade, these techniques should sound vaguely familiar.   

So here in lies the rub.  The problem of causality plagues a lot of variance-driven research.  Causality is, as it sounds, an attempt to clarify whether one action actually causes another.  How scientists study this causal relationship is by looking at how variation in one action (the independent variable) causes variation in another (the dependent variable).  Sounds simple. Until you want to study something interesting, like people. That's because you can't really tell whether the causal relationship goes in one direction or the other. 

Take, for example, teams, morale, and productivity.  You can find that happy teams are productive teams, and conclude that happy people create productive teams because they work better together.  This leads to one set of managerial recommendations around making your people happy.  Or you can find conclude (often more accurately) that productive teams create happy people because people like to do things they do well.  Now you get a completely different set of recommendations around managing people in teams.  Of course, because people aren't dumb, the causal relationship goes both ways many hot teams are caught in the positive feedback of doing well, feeling good, working harder, and doing better.

But back to Google. The company becomes wildly profitable. As they do, they begin to implement a series of management practices that they read or saw that makes people innovative: free food, massages, 20% free time.  Then someone comes along and sees these practices and concludes, naturally, that they are the source of all that innovation.  We all read the same things about Yahoo, Silicon Graphics, and Apple (3M made the 20% rule famous, though in practice when you work 60 hrs a week, 20% is more like your Saturday).

So are these practices really the answer to creating an innovative workplace?
Lashinsky was actually writing about how entrepreneurs were leaving Google to start their own companies because working weekends to create a new application for a search company that has money to burn and applications to spare may not be the best recipe for innovation. 

In other words, no matter what the scientists, journalists, and pundits say about their "innovative" management style, the entrepreneurial rats at Google who are leaving the ship see something else.  Something that is missing from the new and successful Google. Something they would like to return to in order to be successful themselves.

And think about it–did the haute cuisine dining room precede or follow Google's rise to dominance in search?  If it did, we would be seeing more VCs insist their first $1M in goes to building the four-star cafeteria.

I worked at Apple in the early 1990's. When I got there we were at 55% gross margins and when I left we were at 15%.  Was I the cause?  I hope not.  Either way, I got to see how Apple's "innovative" management decisions in the 1980's came back to haunt them in the form of impossibly-high manufacturing costs, impossibly-scalable perks (Friday beer busts, morning bagels, massages, first class travel and accomodations), and impossibly-complex product proliferation.  Three years later, I wandered the halls of Silicon Graphics and saw the same decisions waiting to bring them down. 

The next time someone points out an organization's innovative innovation practices, think hard about whether those practices were there when the real innovation took place.  Then think about where you've seen them before, and what happened to those companies.