TFH Flashback: Self Reliance
May 5, 2014 Timothy Prickett Morgan
People have been buying commercial applications for so long that it is sometimes hard to remember that the System/3X and AS/400 minicomputer businesses were dominated by customers who, by and large, wrote their own software rather than buying it off-the-shelf from a third-party company. Even at the dawning of the commercial Internet era in the early 1990s, the AS/400 community exhibited the kind of self-reliance that we see mostly from hyperscale Web applications these days.
Just for fun, we thought we would take a trip down memory lane and show you what the OS/400 application software market looked like just as IBM was preparing to divest itself of its software assets during its fateful 1993 transition year. That was when IBM decided to become a platform maker and services provider and to partner with other application software providers at the dawning of the ERP software era. Big Blue made plenty of money in the ensuing years, but it may regret not having built a bigger application software business. Anyway, here is the story as it was 21 years ago this month:
When it comes to application programs, AS/400 customers are a very independent lot. About three-quarters of the core financial programs running on AS/400s in the United States have been written by in-house programmers. Some of these accounting programs are direct descendants of applications originally developed for System/3 minicomputers. Others were brought forward from System/36s and System/38s. And a small number of these programs have been written from scratch on AS/400s.
The popularity of homegrown software is in part the result of necessity. Companies that first installed midrange computers ten or fifteen years ago really had no choice but to write the programs they needed. But times have changed. There are literally hundreds of vendors who offer core financial applications. Most of these programs meet the basic requirements of business and cost less than custom software. However, with so many ambitious suppliers in the market, leadership is still up for grabs…in part because IBM has backed away. More than that, IBM has turned over to independent software companies the marketing and maintenance of a number of program products it once controlled. Among these packages are core financial applications for distributors (DMAS), manufacturers (MAPICS) and construction companies (CMAS).
The most plausible explanation for IBM’s failure to offer basic, generic applications is that its management has come to believe that the business is intrinsically unprofitable, at least for a company with IBM’s mass and cost structure. Maybe IBM is right. Although the AS/400 applications business is big–hundreds of millions of dollars every year–it is tough to say whether the business is profitable. Of the big software suppliers, only System Software Associates gets most of its money from sales to AS/400 shops; the remaining large software houses are primarily in other market segments. Though SSA’s BPCS MRP and accounting software is among the most popular off-the-shelf packages for AS/400s, SSA has come under financial pressure in recent months. Computer Associates, a big player in mainframe software, also has a large AS/400-related business, mostly from its acquisition of Pansophic last year. But it is hard to say just how big CA’s AS/400 business is, and impossible to find out what it brings to the bottom line. Software 2000, Lawson and J.D. Edwards are smaller, private companies that keep their financial results to themselves.
Granted, IBM has had a tough time competing against these vendors. The independents are quicker to spot customers’ emerging needs and very good at turning their observations into better and more flexible AS/400 applications. The success of the specialized players isn’t surprising in various vertical markets, but when it comes to programs that could be sold to the vast majority of AS/400 customers, IBM really doesn’t have much of an excuse. Not only has IBM somehow failed to provide superior basic software (on which other vendors could build specialized programs), it seems to have underestimated the importance of cheap, attractive applications programs to its sales effort.
IBM likes to think, for instance, that it can eventually move the thousands of die-hard System/3X users to AS/400s. But there’s no way that System/3X customers can afford IBM’s (or anyone else’s) rates to do a full-blown application conversion. Rewriting applications can not only cost more than buying standard software, it can actually cost more than buying new hardware! But the old applications work and customers are used to them. So there has to be a very dramatic reward for moving to new, standardized accounting software. And it has to be obvious, the way the use of a spreadsheet package is an obvious improvement compared to an adding machine, pencil and paper.
All the independent software houses want to create accounting programs that pretty much sell themselves and, technically speaking, quite a few of them seem to have done a very good job. But there is a big difference between writing software and making it a household word like Lotus 1-2-3, Excel, WordPerfect or Windows. IBM could do this by either creating the programs or putting its name on software licensed from one of the independents. IBM could also help stimulate the rise of a number of software houses by more vigorously promoting their products and at the same time persuading the vendors to compete harder by slashing prices. Whatever the route, at the end of it AS/400s would be more attractive to current customers and much more alluring to businesses (such as those still running System/36 machines) that have considered migration to an AS/400 and then deferred the move.
Even without IBM’s help, the rising costs of applications maintenance and the constantly improving work of the commercial software houses have had an impact on the base. A growing number of AS/400 customers that must consider yet another rewrite of their homegrown accounting software have decided that it might be better to adopt a standardized program instead. If not better, than certainly easier. In addition to the direct savings, these customers cite another benefit: They need all the talent of the in-house programmers to build the special applications that simply cannot be found in the market.
We consider this evolving attitude a sign of maturity. Enlightened customers now understand that self reliance doesn’t always mean ignoring sound commercial software. Most often, it means making independent decisions about what to buy and what to build.