Under Control


Over the years, I've been involved with nine software-based companies. Some of them did not succeed in the marketplace, but the technical stuff came in within budget. By that I mean no overruns of more than 40 percent. One advantage I had is that I ain't no software guy. My background is physics. Tough to know everything about nothing. Ah, but I digress.

In any case, I think my experiences with software can be applied to a lot of different businesses. To my mind, the secret is to start at the top and cycle down. The actual software program coding is the last thing done. The first task is to write a sales brochure. This establishes the basic offering of the product. The brochure is what gets buy-in from the engineers. Everyone must sign off on the project goals and product definition. The most difficult sell is always the internal one. If you don't make that sell, you'll encounter the enemy. The enemy is anarchy. The brochure answer the age-old question, "What business are we in?".

Then comes the "bible." a 200-page tome in which you describe the details of the implementation, including types of work stations, operating systems, and platforms. The author is usually a high-IQ landscape artist with some moss on the north side. The bible is written from a marketing viewpoint, not a sales or technical outlook. It is, however, written for them implementers and quality people. It is not done by committee, but lots of interaction is needed to get it done. And it can take more time to do than actual coding. Although not a perfect document, it must have no fatal flaws.

Everyone read and sign off on this internal document. It must be generic enough to meet all of the needs of the user, without constraining the demands for new features on the part of sales. From that point on, all questions can be met with the same answer: "Is it in the Book?". No extras or changes are allowed once the project is underway, unless the changes fixes a fatal flaw.

I know of a company that always has three projects underway and never finishes any. Their product vision is so short term that the chosen technology is invariably obsolete in 18 months. And anarchy rules there, with the engineers blackmailing the management in order to achieve their own nefarious ends. Thus a "new" project starts every six to nine months, incorporating the latest, etc.

Managing an organization is a classical engineering problem. It is not an art form. Marketing input, brochure, bible, and implementation - in that order. Feedback is frequent and substantive. Organization is flat and loosely coupled. Teams of two to five people are assigned components of the bible.

Management is achieved by watching resources, not objectives. In other words, staffing levels are a given, and the project prices must conform to them, not vise versa. Meanwhile, the moss-covered grand old man circulates, and the while giving forth little gems of wisdom. Baseball teams built around stars seldom win the pennant.

Let's take an example. Assume the assignment is to do a networked financial transaction package. The first offering is a desktop blister pack. The Bible is written and calls for use of three-man teams. The segments include user entry, database, optimizer, communications, reports, screen, and management. Each is treated as a product for sale. Each has testing and quality procedures. Each segment is defined well enough to be subcontracted to outside suppliers. About 30 percent of them will be. The secret to software management is to manage the process, not the project. The first pass should be a prototype and treated as a throwaway. As for the engineers, give them all the tools and tell them when they are done.

As each piece is completed, it must be signed off on by the moss-covered guru, marketing people, and focus groups. The metrics of implementation must be rigorously adhered to. Early field testing is done by the internal users and selected forgiving customers.

As I have said before: software is someone else's idea of what you want to do. Test and maintenance is 80 percent of the task. The problem is, small companies have too many chiefs and not enough Indians, while large companies have too many Indians. Discipline, small teams, and attitude will make it happen.

As appeared in Manufacturing Systems Magazine February 1995 Page 16

Manufacturing Systems Magazine Articles
References - Table of Contents

Send mail to rmi.info@barn.org for more information.
Please send mail to webmaster@barn.org regarding web site structure.
Copyright © 1996-2002 R.Morley Inc. All Rights Reserved

R. Morley Incorporated
614 Nashua Street, Suite 56
Milford, NH 03055-4992 USA
Tel: 603-878-4365 FAX: 603-878-4385