To buy or be bought

To buy or be bought
Dec 29th 2004 SAN FRANCISCO From The Economist print edition

BUSINESSS
oftware takeovers

Everybody agrees that the software industry is starting to consolidate. But who will buy, and who will be bought?

TWO big software takeovers at the end of 2004 have prompted bosses in the industry to ask themselves whether they should buy or be bought in 2005. On December 13th, Oracle, the world's second-largest software firm, prevailed in its long struggle to buy PeopleSoft, a rival, for $10.3 billion. A few days later, Symantec, the largest vendor of security and anti-virus software, agreed to buy Veritas, which sells storage and data-recovery software, for $13.5 billion in shares.

These deals have “put blood in the water”, says Jonathan Schwartz, president of Sun Microsystems, a large vendor of hardware and software which is one of the companies that could find itself playing the role of either hunter or prey in the coming year. The danger, he says, is that there could be “mass confusion between activity and progress” as firms leap into “awkward combinations” for fear of being left behind. For instance, the logic of combining Symantec and Veritas, which have very distinct businesses, is not obvious. In the large and complex software universe, which mergers do make sense?

The first thing to realise is that consolidation is happening overwhelmingly in the market for business, rather than consumer, software. In the consumer market, despite the best efforts of the antitrust authorities, Microsoft continues to dominate operating systems and office applications.

The market for business software breaks down into two broad categories. The first is applications. These are programs that automate corporate functions such as supply-chain planning, accounting, marketing and payroll, and that sit at the very top of the “stack” of software layers that run corporate data centres. The second category, infrastructure, includes all the layers underneath, from “middleware” that acts as software glue between systems, down through various management and security layers to databases, storage software and operating systems.

A universe of opportunity

The applications category is already much closer to the end than the beginning of consolidation. That is because the merger of Oracle, which is primarily a database company but is also the world's third-largest applications vendor, and PeopleSoft, the second-largest applications firm, will produce a quasi-duopoly with SAP, the market leader. A few niche players, such as Adobe or Autodesk, may well remain independent, but many medium-sized applications firms, such as Siebel, a struggling maker of sales and marketing applications, look vulnerable. Even so, there are not many takeover targets left.

This means that most of the merger action will be in infrastructure software. Paul DiNardo, a technology specialist at Credit Suisse First Boston, an investment bank, likes to picture this industry as a galaxy with two dominant stars, Microsoft and IBM, each orbited by large and small planets—other firms with which they have technological affinities. Those orbiting Microsoft work mainly with its operating system, Windows, and its associated software, while those around IBM prefer non-proprietary “open-source” software, such as the Linux operating system.

One possibility is that aliens could invade from another metaphorical galaxy. Three possible invaders are Cisco (dominant in the networking galaxy), Intel (in semiconductors) and Dell (in hardware assembly). This scenario is not very likely. Dell knows that its success is due largely to its focus on manufacturing efficiency, not software. Intel is battling to keep its core market. Cisco is acquisitive, but its equipment sits mostly at the edge of corporate data centres, not at the centre.

It is more likely, therefore, that consolidation will occur within the software galaxy. One scenario is that the stars will gobble up planets in their orbit one at a time. For the small planets this is already happening—IBM has bought several small firms in recent months. Some medium-sized planets might also make good additions: IBM could buy Novell, a big Linux supporter in which IBM already owns a stake. But too much consolidation could cause concern among regulators, customers and partners. Microsoft briefly flirted with bidding for SAP (which is expanding from applications into infrastructure with its NetWeaver platform), before deciding that it would cause too much trouble.

Another scenario is that the biggest planets could try to become stars by buying medium-sized planets. Both Hewlett-Packard (HP), which is primarily a hardware-maker but also sells infrastructure software, and Oracle, whose core strength is still in infrastructure, are sporadically said to be talking to BEA Systems, a Silicon Valley neighbour that competes with IBM in middleware. HP has, however, barely digested the huge acquisition of Compaq, another hardware-maker, and is facing pressure from investors to split up and become more focused rather than to expand further. Larry Ellison, Oracle's boss, has suggested that integrating PeopleSoft will keep Oracle busy for some time.

That leaves a fourth scenario, in which the medium-sized planets combine. For instance, Steven Milunovich, an analyst at Merrill Lynch, thinks that it would make sense for Sun Microsystems, which sells an operating system called Solaris that is similar to Linux but competes with it, to buy a dedicated Linux vendor such as Red Hat or Novell. Similarly, EMC, a giant in data storage that has bought storage-software firms such as Legato, VMware and Documentum, could look for more targets. And even firms that do not have much in common could team up for the sake of scale, as Symantec and Veritas did.

What seems certain is that software consolidation will continue for years. Even in hardware, where the process has long been under way, it goes on still. IBM is selling its personal-computer business to Lenovo, a Chinese company. Gartner, a consultancy, predicts that three of the top ten PC-makers will be gone by 2007. The software industry is still behind hardware in this respect and so has much further to go. Expect plenty more blood in the water.

Copyright © 2005 The Economist Newspaper and The Economist Group. All rights reserved.