So-called systems integrators such as Accenture and IBM may benefit most from the growing emphasis on service-oriented architecture
Silicon Valley is crazy about its catch phrases. And when it comes to software, there are plenty to throw around, from tipping points to paradigm shifts to game changers. The thing about those terms and the undercurrents they describe is that they don't benefit the software vendors or the investors. It's always another set of companies higher up the food chain that figures out how to better capitalize on new technology.
And even so, it's often only a temporary win. Eventually, new predators come along and figure out how to gain an even bigger advantage.
We're seeing this played out now in the business-software market. The next "game changer" in the industry is what's known as service-oriented architectures (SOA). Companies such as SAP (SAP
) and Oracle (ORCL
) are pouring billions of dollars into creating such platforms.
CIRCLE OF LIFE. But it will be systems integrators—companies such as Accenture (ACN
) and IBM (IBM
), and Indian companies including Infosys, Satyam, TCS, and Wipro—that benefit big from that investment. The integrators will stay fat and happy until they too are usurped by a new breed of vendor.
It's the life cycle of enterprise software, and we've seen it before. It happened in the second half of the 1980s, with so-called manufacturing resource planning—software that helps companies with business and manufacturing planning. MRP II, as it's known, became a billion dollar software industry, dominated by now-defunct companies such as ASK Computer, Cullinet, and Dun & Bradstreet Software.
The real powerhouses, however, were the hardware companies. For every dollar of software sold, Digital Equipment, Hewlett-Packard (HPQ
), and IBM made several times that amount on hardware, operating systems, networks, storage, terminals, peripherals, and services.
WAKE UP CALL. The hardware vendors' advantage was short-lived, though, when buyers started building their information strategies on databases from Informix, Ingres, Oracle, or Sybase (SY
), which they saw as integration hubs for all their software. The hubs never really caught on. Soon, the whole MRP II market ran out of steam and buyers sat waiting for the next "disruptive technology."
That came in the 1990s and the early part of this decade (roughly 1991 through this year) with the rise of so-called Enterprise Resource Planning, or ERP. Buyers woke up to find that their IT landscape consisted of thousands of applications. Information about customers, inventory, suppliers, assets, and employees was stored in a profusion of data silos. The first wave of ERP software promised to consolidate all of these disparate applications and databases into a single software system.
The rise of ERP was aided by two business trends: executive hysteria over the need for business process re-engineering (redesigning and automating business processes to create a more flexible and agile enterprise) and the industry's shift from mainframes to client/server technology. This shift occurred because mainframes were perceived as barriers to reengineering business processes. The systems were expensive and difficult to use, which runs counter to having a flexible and agile enterprise.
FULL SERVICE. While ERP grew to be a $25 billion software industry, the real winners for the first eight years were IBM and the Big 8 accounting firms that quickly built consulting practices in business process re-engineering (BPR) and ERP implementation.
The integrators sold the BPR engagements, advised executives on selecting the software that would help reinforce the new business processes, and finally closed the loop by actually installing the hardware and software that the executive team had asked them to help choose.
ERP implementations turned out to be very lucrative. For every dollar SAP generated in licenses, the integrators made at least five to 10 times that during the resulting multi-year engagement. Selling ERP projects ensured the fast track to partner.
SAVING THE LEGACY. Which brings us to the present day, and the dawn of a cycle I expect to last at least through 2012, as SOA spawns the ERP replacement market. SAP and Oracle are constructing their own service-oriented architecture—software systems that allow customers to access and reuse logic and data that already exist in a system, and to make it available to other applications through Web services.
To put it in plain English, SOA could free customers from forced software upgrades every two to four years or save them from having to do wholesale replacements every eight to10 years by letting users access and reuse the legacy systems.
While the ERP vendors didn't create the market for Web services or SOA, they're best positioned to stimulate market demand because of their vast customer base. It's going to take a lot of capital and risk, though, to meet that demand. In a recent call with analysts, SAP CEO Henning Kagermann said his company will have 500 Web services available this summer. While this is a good start, company executives estimate they may need as many as 30,000 Web services across all of SAP's software.
CERTIFIED SALES. The promise of SOA is luring a wide range of technology vendors. Cisco Systems (CSCO
), EMC (EMC
), HP, IBM, and Intel are building SOA appliances for networking and data management. Others are building tools for managing SOA performance and security. And thousands of software companies are approaching SAP and Oracle about becoming "certified" in the hopes of selling into their large customer bases.
While the technology vendors see the marketing advantages of SOA, the integrators are eyeing the money. Here's how it will play out: ERP customers will begin to evaluate the new technologies from SAP, Oracle, and others. They will likely ask their preferred integrators to help them redesign their architectures to reflect proposed changes in their business processes. The integrators will advise them on issues relating to architectures, security, governance, performance, network requirements, and change management.
IBM'S EARLY LEAD. Soon, consultants will be arriving on-site by the busload for multi-year engagements. They will develop repositories of Web services and composite applications (the latter will be software that adds, modifies, or extends a business process). Looking at history, the primary difference between the SOA and ERP booms is that many of the newest beneficiaries will be India's off-shoring firms.
The smart integrators are starting to hire now. They don't want to be caught short-staffed when the boom hits. During the ERP craze, there was a dearth of available talent. IBM has the early lead, having completed over 1,800 SOA engagements for more than 1,000 clients. In mid-July, Accenture announced plans to invest $450 million over a three-year period to enhance its SOA capabilities.
Here's where the scenario may turn into a nightmare: Within a few years, companies are buying less software from SAP and Oracle. While they continue to pay maintenance, they now have tools for creating their own applications to sit on top of their legacy enterprise systems, or they continue to use low-cost developers to build custom software.
BOOM AND BUST. The irony is that SOA, Web services, and new industry standards could tip these legacy systems into commodity products. Soon customers begin complaining that it costs too much to maintain the legacy ERP systems at the bottom of the architecture.
By 2010, word gets out about a new type of application built using model-driven development tools. The new software comes with built-in Web services, intuitive user interfaces, embedded analytics, and easy-to-build applications. Venture capitalists rush in to fund the new companies, setting off a gold rush not seen since the Internet bubble. It has happened before.
Richardson is chief research officer at AMR Research, a Boston-based research and advisory firm specializing in enterprise software
