Watch out for those cow paths

The cow path is still out there, begging to be paved, and in fact some IT practitioners are looking to milk clients in the process. Thus, clients have a problem. They know two things: that they have a business practice that is causing them pain and that computers can help.

Unfortunately there are three things to know. The third is how to make sure computers really do help.

Humans have a nasty habit of letting themselves dig little ruts in their brains. Have you ever made a mental note that, one morning on the way to work, you are going to stop at the shoe repair place and drop off some shoes for repair? Have you ever gone out the door, shoes in hand, and promptly driven to work in exactly the same way you have for the last zillion weeks, totally forgetting about the shoes? Work processes are like that. Computers are supposed to make whatever it is we do now faster and more accurate. This is the equivalent of putting a jetpack on a cow — it may be fast, but it’s still a cow. Changing the fundamental nature of that business process is hard to do when you are on auto-pilot.

IT people have a problem too. Out of deference to the client, and a genuine ignorance of the client’s business, we believe that what they are telling us is true and the way they want the work process automated is accurate. They aren’t lying about what goes on currently, but they probably have a blind spot with respect to improvement. We need to listen to the adage that you need a computer to really screw something up. An automated bad process is a nightmare.

Blind – or wildly unscrupulous – IT people have charged clients a lot of money for systems that never worked. The code and database design might have been great, but it was applied to a bad business practice.

How do you avoid this?

Don’t accept work when you know that the business practice is loopy. For example, it is pure rework to write a system that validates data that has already been scrutinized somewhere else.

Don’t just say, “Whatever you say – you’re the client.” Tell and then show clients that the business practice has flaws. Suggest ways to improve it.

Learn the business for real. Too many IT people leave one industry, go to another, and say, “This is all the same – data, process and donuts.”

Mock up the process. Use a low-tech tool to test process ideas. If you are writing a multi-user database, quickly cobble something up on a PC package and see if it works. The very act of creating the process will illustrate all the assumptions you made and errors you introduced.

Find your own blind spots. What assumptions have you made?

Obviously, you won’t know what assumptions you’ve made, because if you knew they wouldn’t be assumptions. Describe what you are doing to another party whose interest and knowledge in what you are doing is limited. (The problem is that you may bore the hell out of the person. I suggest coercion and coffee to keep the subject awake.)

What you are looking for from your friend is a statement akin to: “You can’t put jetpacks on cows — that’s crazy!” Then you will have found your blind spot.

Ford is a Vancouver-based systems consultant. He can be reached at