Herding Cats for IT Governance
Take an architectural approach to gaining agreement on technology priorities
by Tom Pardee
Posted November 28, 2003
One of the most common frustrations for architects is the difficulty in obtaining project approval and then keeping everyone on track to implementation. This problem, almost universally described as "herding cats," can only be avoided for projects that have a short term ROI, are mandated by regulation, have a C-level sponsor, or are critical to avert impending doom. The frustration is compounded when the same investment that was rejected one year becomes a cost-is-no-object, do-it-yesterday project a year later.
But our difficulty in getting agreement is more than management's failure to listen or be reasonable. It is the absence of a suitable communications system. To the extent that architects can support the decision makers with a more rational approach, they will help their enterprise and themselves.
Communications Theory
By now most enterprises have adopted cross-functional governance committees and related structures to allocate and control IT investments. These are filling the void left after decision making was wrested from IT at the beginning of the 1990s, but they do not address the fundamental problem of how to bring clarity to choosing between projects or deciding an investment is worthwhile.
Many observers have cited the lack of expertise on the governance committee as a primary problem. They argue that executives were successful because of expertise in sales, operations, or some other discipline that is primarily concerned with customers, staff, partners, and business imperatives. In fact, governance committee members come up to speed quickly if their role and the process are well defined. The more pressing problems lie in the nature of the assets to be managed and the complex proposals that promise to increase their value.
The first challenge is conflicting goals in the enterprise and the complex data, application, and technological infrastructure that embed them. The difficulty of rationalizing these assets is multiplied because they are hidden in a virtual world of intellectual capital. Before commissioning a major cross-functional project that targets data or application architecture, committees must be certain that they understand how to reconcile these hidden tensions.
Requests to upgrade infrastructure are even more of a challenge because the technologies are outside of the training, expertise, and even interest area for many general managers.
To this day, IT managers and architects approach governance by creating voluminous requests that describe everything from the environment and history of the problem, to data structures and performance metrics, to financial models purporting to justify the investment in terms of ROI. Then we throw it over the wall into the decision process like a Hail Mary pass—a desperation play that has only a small chance of succeeding.
Such a decision process can be characterized as involving a poorly prepared information-processing system (the governance committee meeting) and complex messages (our project requests). These characteristics are exactly the opposite of what modern communication theory recommends: a well-prepared system and simple messages.
The state-of-the-art model for good communication in a complex environment like ours is based on carefully selected metrics in an intuitive analog framework, taking advantage of our natural ability to process such information. Messages are filtered and transformed, and then presented as readings, comparisons, or trends. A decision maker compares three numbers in a series, recognizes that a graph is headed upward, or sees that the blue slice of pie has grown in relation to the red slice. Such a system is called a dashboard.
As there is no avoiding basic business complexity, a dashboard typically contains a number of different graphics for different indices. In addition, the dashboard allows drill-down capability so the decision maker can dig into the information supporting any index value.
The Governance Dashboard
An architectural dashboard for governance would support resource allocation decisions, presenting relevant information about the project in the context of real enterprise values. Like any dashboard development project, there are three parts: finding the right graphical paradigm, getting agreement on the metrics to be captured and shown, and presenting the actual data against this context.
Part 1: Perception Charts – A highly effective presentation for the kind of information architects need to present is the Perception Chart. These are square graphic plots divided in quadrants; the x-axis and y-axis show useful values, and the subject of the discussion is plotted according to its ratings for those values. When several options are plotted together, their relative values jump out at the viewer.
Back to top
|