An Architect Looks at Organization Design

The first in a series of articles by an architecturally-trained designer on the shaping of organizational change MOST OF US, when we fly on an airplane, look at what the cabin attendants and pursers do and think to ourselves, I could do that. Yes, there are rules and procedures to learn and a manual to master, but nothing I could not do if I spent a little time at it. Our reaction to the work of the pilot, however, is altogether different. After years of training, we might be able to fly an airplane, but we might never become very good at it. There is an issue here of basic skill, a question of aptitude and ability that goes well beyond simply learning the right sequence of procedures. Most managers view organization design as being far more like the work of cabin attendants than that of pilots. We believe we all could do it if only we put in the time. After all, we spend most of our working lives in organizations. We are "there" every day. We know all the twists and turns and hidden spaces that a visitor might miss or might not recognize, just as we know all the idiosyncrasies of the house or apartment we live in or the car we drive. In practice, however, designing organizations to meet new competitive challenges or to accommodate new technologies or to address new markets is often an exercise in frustration. Analytically-gifted managers devise new product/market strategies, compute a division's real cost of capital, or lay out the process flow for an assembly line -- if not with perfect ease, then at least with the confidence that comes from tested skill. But designing organizations -- sketching out which business units belong where, which solid or dotted lines connect them, which information flows cement the linkage -- is something else again. The familiar tools and approaches do not fit. Trying to bang them into useful shape does not work. And no obvious substitute lies ready to hand. This, of course, is because organization design -- not establishing specs for an organization ("let's create a matrix structure," say), but actually moving from those specs to a design that embodies them -- is not, at base, a work of analysis. Analysis, however, is what most managers are good at. It's familiar. It's what they've been trained to do. It's what they know how to do -- and to do well. In general terms, analysis is what all of us do when we take an alarm clock apart to see what's inside, why the hands move, why the alarm bell rings. But when it comes to putting all the scattered pieces together, to making sense of the most appropriate "fit" among them, most of us have trouble. There are skills we lack, for what this sort of task requires is not analysis but synthesis and integration, a sense -- literally and figuratively -- of spatial relationships, of offsetting weights and balances and masses, of movement and proportion. From our earliest school years, most of us have been taught how to take things apart in order to study them. Even if we do not know precisely what steps to follow in a specific case, we usually have a pretty good intuitive feel for the order in which to try them. We can usually recognize and interpret the evidence that tells us whether we are on the right track. And if instruction sets are available, we can usually follow them. With synthesis and integration -- that is, with design -- it is different. How to get from here to there is not so easily taught. Intuitive judgments are not so reliable. And instruction sets are not much help. Like good architects, what we need, instead, are an awareness of context and a feel for what good "fit" might look like. Does creating a particular reporting relationship overload one part of an organization? Does creating a particular information flow weaken another? We also need a willingness to tinker with ideas and possibilities until they lock into a suitable shape. What we're after is not the shortest possible line straight to the heart of things, but an open-ended process of experimentation, an iterative sketching-out of rough solutions. …