I've been worried about the health of our domestic reseller channel for over three years. Many of the dealers who sell the bulk of our products today, in an expanding economy, are marginally profitable or losing money. As I pointed out last March in the memo ``Are we bankrupting our dealers?'' (copy attached), Autodesk's pricing policies may have contributed to this situation. If the economy contracts even a modest 10%, the probability is extremely high that many of our dealers would be forced out of business or into other sectors of the market.
If this happens, and Autodesk is not prepared to cope with it, it could be a catastrophe of the greatest magnitude for our company. Reseller distribution has been absolutely central to Autodesk's entire business concept since 1983. It is our dominance of the reseller channel that has made it so difficult for competitors to assail our market share. (The fact that we react so quickly to any attempt by competitors to recruit our dealers is evidence that we recognise this fact.) The very composition of our products and the structure of our company have been built around the dealer channel. Our promotional campaigns and sales literature reflect it. Devolving installation support, customisation, consulting on hardware configurations, selection and integration of application packages, informal training, and, most importantly, post-sale support, upon our dealers has been essential in maintaining Autodesk's very high profit margins over so many years. In a very real sense, we have offloaded many of the traditional cost centres of a CAD company onto our reseller organisations. The price we've paid for doing this is the risk we now run if those businesses begin to fail.
Unfortunately, the dealer viability crisis not a problem which Autodesk has much leverage to solve, as far as I can see. If, as I believe the market has demonstrated unambiguously, the value of AutoCAD to a customer is between $2,000 and $2,500, then further price increases will only further consume the resellers' margin in forced discounts, reduce overall volume, and contribute to market share erosion in favour of lower-priced alternative products.
Although Autodesk has contributed to its dealers' woes, the fundamental problem is not of our making. In the early days of AutoCAD, it took an experienced dealer simply to obtain the odd collection of hardware that was required to lash up a machine suitable for running AutoCAD, get it all to work together, and tweak the system so it delivered acceptable performance. Many of the peripherals required by CAD users, notably professional quality digitisers and large drafting plotters, had essentially no retail distribution channel at all. And, early AutoCAD was no great shakes when it came to ease of installation or the initial learning curve for a new user to become productive.
In such an environment the AutoCAD dealer added real value that was readily perceived by the customer. This is usually the case in emerging markets--it closely paralleled the development, a few years earlier, of businesses which configured and sold word processing systems based on CP/M machines and WordStar. The eventual demise of that business provides insight into what may be happening now in our reseller channel.
As the market has matured and the price-performance of computers and peripherals has skyrocketed, a CAD system is no longer a odd collection of exotic hardware that can be assembled only by a skilled technician. Instead, today one can order everything needed to assemble a high-end CAD system over the telephone, toll-free, and expect that when the boxes arrive it will take no more than a careful reading of the instructions to get everything working. Architectures such as the Macintosh and SPARCstation where component integration has been carefully thought out contribute to this trend.
In addition, a customer who previously turned to his local dealer as the only source of knowledge about AutoCAD can now read any one of dozens of books about the product, attend courses at a local school, call a 900 number support service, read an article in Cadalyst or Cadence, post a question on CompuServe, or turn to a fellow AutoCAD user for assistance either at a user group meeting or informally--our very success has engendered a proliferation of alternatives to the services originally provided only by dealers.
I wish these fundamentals were different or that I could find a fix for the problem. Our dealers have worked very hard for us for a very long time, and their efforts have been largely responsible for our success. But if the dealer channel is becoming nonviable, it is incumbent on Autodesk's management to anticipate that eventuality and have plans in place to cope with it should it occur.
Editor: John Walker