« January 2007 | Main | July 2007 »
April 19, 2007
Bill Boquist
Enterprise Architecture within a Social Ecosystem
I'd like to offer an observation that bothers me as I think about the field of Enterprise Architecture: as we engage our business partners and develop our models, we tend to treat the complexity of humans beings as information processors as a part of the context surrounding our work, rather than a topic within its scope.
If we reduce the enterprise to its barest essentials, it is nothing other than a device for acquiring, structuring and acting upon information. Therefore, if we seek to optimize its operations, we must be willing to dedicate a time slice to understanding the motives, strengths and weaknesses of the actors who are the principal agents of information flow. We're not doing that today. Our tendency to leave the topics within this realm relatively untouched manifests itself in two primary ways. First, in our own day-to-day work, we take it upon ourselves to navigate the political ebb and flow as best we can, but invest little or no time pondering the possibility that with time, thought and effort, we may be able to engineer a lot of the politics out. Second, consciously or not, we frequently operate as though the actors who will execute the work of the business are "black boxes" about whom we can know little.
Both of these habits undermine our effectiveness. We need to include a deeper understanding of the people who operate or otherwise interact with the enterprise as one of the strengths we bring to the table, and factor that understanding into our work. If we realize that the building blocks of any enterprise are its people, then I think the field of enterprise architecture can fully succeed only if we develop (and design with) a fairly profound understanding of the strengths and weaknesses of people as economic and social agents. Then, we can create operational specifications for the enterprise "machine" that use the strengths fully and compensate for the weaknesses.
To illustrate with a simple analogy, we expect any architect of bridges to know that steel expands and contracts as temperature rises and falls, that it oxidizes much more quickly in the presence of salt, that it loses structural strength well below its melting point, that it is strong only in tension when woven into cable, etc. We don't expect him to be a metallurgist, but we do expect him to be conversant with the fruits of the metallurgist's research, and to design the structure accordingly. I would argue that we should expect similar expertise from an enterprise architect - not that she should be a psychologist, but that she should have a fairly good understanding of the ways in which human cognitive and emotional functioning departs from the ideal in predictable ways. Then, she can design the enterprise to be an entity within the social ecosystem that is more rewarding to all stakeholders (employees, customers, suppliers, etc.). Fundamentally, an enterprise design is nothing more than a blueprint for structuring people's interactions to produce more value than the same set of people would produce without the structure. So, shouldn't the design be driven by the functional properties of the actors?
Here are a few specific instances of ways in which social scientists have long known that the human mind tends not to work precisely as we might expect:
- People tend to find it difficult to visualize a future that is different than the past. We're natural extrapolaters. For example, when the housing market has been going up for a few years, almost no one can conceive a future in which the median house price will be flat or falling. Similarly, overbuying when times are good and overselling when they are not has plagued the stock market for decades.
- People tend not to periodically re-test the assumptions on which they are basing their behavior to determine whether or not those assumptions still hold. Encyclopedia Brittanica did not properly adjust its business model when Microsoft Encarta debuted, because they continued to assume that all "serious" reference would be published in paper form. They had to declare bankruptcy a couple of years later.
- When preparing to reach a decision, people tend to factor in only a subset of the relevant information, because the full range of variables exceeds the "seven plus or minus two" that they can deal with concurrently.
- People tend to acquire imaginary "property rights" in places and roles they have occupied for a long time. This gives rise to fief building, infighting and other maladaptive uses of enterprise property.
- People tend to perform better when moderately challenged, especially by novelty, and less well when bored.
- Mathematical diagnostic models derived from expert's judgments frequently outperform the experts themselves because the models eliminate the "noise" in human thinking.
If we know of these (and many other) tendencies, we can design to mitigate them explicitly. If we don't, we can only hope they don't impinge too greatly on our ability to produce value by applying structure. From my perspective, it is only responsible that we understand the properties of the "material" we work with as enterprise architects.
Posted by GEAO22 at 07:08 AM | Comments (3)