Integrated IS Plan #8: Information architecture (first appeared in InfoWorld)

Like Tweet Pin it Share Share Email

A CIO of my acquaintance once described his priorities for the new megasystem his developers were busily constructing. “What I want is the database,” he said. Waving his hands disparagingly, he added, “I don’t much care about the applications that feed it.” Guess what? His system never got built, and he didn’t last two years in his job.

We’re continuing to develop the technical architecture section of our integrated IS plan. This week we home in on the information layer. The techniques for managing this layer are well understood. Instead, let’s elevate the discussion (elevate in the sense that snipers like high places). Our goal this week is to place information in the proper context. We begin by avoiding the mistakes of the aforementioned CIO, realizing that although information is the center of our universe, applications drive the business.

We made a horrible mistake when we changed our name from electronic data processing (EDP) to management information systems (MIS) back in the early ‘80s.

When we were EDP, we did something valuable – we processed something, and as we did so, we automated manual processes.

MIS managed information. Even worse, we declared that our purpose is providing information to managers. Helping employees do useful work became a byproduct.

We would be much better off calling ourselves “process automation systems.” We got offtrack because of database management technology. With the advent of the DBMS, we changed how we designed systems. We put information in the center of design. Information, we realized, is more stable than the programs that make use of it.

Next we figured out that because information is the heart of our designs, it must be at the heart of the enterprise. So far so good, but then we left the halls of reason and jumped to the notion that it’s information, not processing, that delivers the most value.

Take a fresh, hard look at this. IS delivers the bulk of its value through process improvement: lower unit costs, reduced cycle time, and increased accuracy.

This is just as well. If we really do think information is the point of it all, our efforts are way out of whack with the company as a whole. About 80 percent of an average company’s information is unstructured. (I’ve run across this estimate several times, and it passes the “feels right” test, too.) It’s text, voice, and pictures. A simple-minded feller might figure that if information is the point of it all, and 80 percent of all information is unstructured, well then 80 percent of our efforts should be devoted to the management of unstructured information. They aren’t, of course. Eighty percent of our efforts go to managing alphanumeric data – the kind we know how to process. Telephone systems and personal computers – the technologies that handle unstructured information – have been the poor stepchildren of IS, not because we couldn’t manage the information, but because we couldn’t process it.

We’ve been able to get away with this so far. No longer, though. Maybe it’s the influence of e-mail and the World Wide Web, but companies are waking up to this deficiency.

Want to know your future? Look at a modern call center. It records every conversation digitally, along with every screen visited during the call. It’s indexed and ready for online retrieval to help call center management assess individual performance. It’s also available for computing sophisticated performance statistics.

Today these systems are closed and proprietary. Tomorrow they’ll store everything in the same document management system that will store scanned images and word processing documents, all linked through a common index. That’s just one example of how you will have to manage and process information in the near future.

Thought your information layer was in good shape? Maybe for today, but you have some serious planning to do.