No-one Makes Their Own Today

Sydney Morning Herald

Thursday September 2, 1999

GRAEME PHILIPSON

One of the big growth areas in software in the last year or so has been what is called customer relationship management, inevitably shortened to CRM. There have also been trends towards supply chain management (SCM), sales force automation, and other types of software such as planning and logistics.

All of these types of software have one thing in common: they are packaged solutions to problems that were, in most cases, previously solved by software written on a one-off basis.

There has been an enormous trend over the past 10 years towards off-the-shelf software, often called packaged software, and away from programs created internally by companies. The massive growth of the ERP (enterprise resource planning) market in the 1990s is one of the most noticeable aspects of this development.

The rise in packaged software has many causes and many effects. It has been one of the key defining features of the 1990s, as important as the ubiquity of PCs, the growth in networking and the rise of the Internet.

The software industry did not even exist until 1969, the year IBM began marketing software and hardware separately. In the following years many software companies came into existence but most of them developed products to help people run their computer systems (this was called systems software or systems utilities) or develop applications (this was called applications development software).

Applications development software was one of the big growth areas in the technology industry in the 1970s and 1980s. It evolved in the late '80s into something called CASE (computer-assisted software engineering).

As I mentioned last week, CASE was the ultimate in applications development: using software to write software. There was much talk in the late 1980s about the ``applications development backlog" and how most computer installations were falling behind in maintaining existing applications and building new ones. CASE, by automating and simplifying this process, would save us all from these problems.

But a funny thing happened. CASE disappeared, because the applications development backlog disappeared. Companies, by and large, stopped writing their own programs. Strange as it may seem now, 10 years ago organisations were still writing their own payroll and general ledger programs. Nobody does that any more: you simply buy a package ``off the shelf".

Packaged software became popular because the ``make" versus ``buy" equation changed. Packages got better, and more tailorable, at about the same time the PC revolution hit. This had another major effect. With the growth of graphical user interfaces like Microsoft Windows, anyone could for the first time write their own little quick and dirty applications, using commands built into spreadsheets and other programs.

In the old days, it took an enormous effort for an end-user just to get a column moved or widened on a report. The user had to go to the internal applications development department, which then had to allocate the resources to rewrite the computer code that determined the column width, and then the code had to be actually rewritten.

It was an expensive, laborious and error-prone process. No wonder the users rebelled and no wonder they got their own little desktop PCs as soon as they could. Add to that the increased performance of packaged software, and applications development went west very quickly.

The first products to be turned into packages were the core applications like financials and payroll. But other applications soon followed. A good example is to observe the changes in data retrieval.

There has always been a need for software that makes it easier to extract information from computer systems. In the 1980s there came into existence a breed of software known as executive information systems, designed to help senior management make sense of all the data in their corporate computer systems.

EIS evolved into business intelligence, and that into something called data warehousing. But these were not applications as such, they were ways of building applications the user still had a lot of work to do. Now off-the-shelf CRM systems essentially do the same thing but they just need to be filled with data.

Soon, no-one will be writing applications except the software suppliers. And with the move towards applications services, perhaps soon no-one will even be running any applications. But that's another story.

Graeme Philipson is an information technology analyst. This column examines the events that have shaped today's computer industry.