IT doesn't matter--business processes do

In this excerpt from their critical analysis of Nicholas Carr's IT article, the authors dissect Carr's view of IT as a commodity, examine the danger of lumping all IT into a commodity bucket, and posit that IT will play a major role in making tomorrow's i
Written by Howard Smith and Peter Fingar, Contributor

CTO View: In a forthcoming book IT Doesn't Matter--Business Processes Do: A Critical Analysis of Nicholas Carr's I.T. Article in the Harvard Business Review (Meghan-Kiffer Press, 2003), Howard Smith and Peter Fingar offer their perspective on the controversial article, "IT Doesn't Matter," written by Nicholas Carr that appeared in the May 2003 issue of the Harvard Business Review. As you might recall, Carr's article incited rebuttals from every pundit and industry executive, as well as some concurring nods on some of his points. In the following book excerpt, the authors dissect Carr's view of IT as a commodity, examine the danger of lumping all IT into a commodity bucket, and posit that IT will play a major role in making tomorrow's innovations happen.

Applications versus Business Process Management
The Issue: Carr's article is about functional applications, not end-to-end business process management, which is the real frontier for business and IT

A central problem in Carr's premises is his use of the term "application" when he writes, "IT is also highly replicable. Indeed it is hard to imagine a more perfect commodity than a byte of data-endlessly and perfectly reproducible at virtually no cost. The near-infinite scalability of many IT functions, when combined with technical standardization, dooms most proprietary applications to economic obsolescence. Why write your own application … for supply chain management when you can a buy ready-made, state-of-the-art application for a fraction of the cost?"

Today's companies are less focused on applications than they are on tying applications together internally and with business partners to automate end-to-end business processes. But even within the realm of applications as Carr describes them, standard IT solution packages don't dictate the processes running on them, that's determined by the businesses that use the packages. Tailoring today's packaged enterprise applications is a common, though tedious, means of making an otherwise packaged solution function as, what Carr calls, a "proprietary application." This is old hat. Paul Strassmann elaborates, "Even the most tightly controlled generic application suite (SAP's enterprise resource planning application) can deliver completely different results for look-alike firms." Even so, the business issues in IT go well beyond the issues of computer applications and on to the digitization of business processes themselves-a trend that's been well underway for the last few years.

For the past fifty years, business "applications" have been all about functional automation. Today, most of a company's business processes are trapped inside software that automates one specialized, and usually inward-focused, back-office business function and its activities, e.g., the inventory system or the order entry application. Companies began to use IT tools such as enterprise application integration (EAI) to tap the specific business processes trapped in functional applications, which were often buried in ERP suites, and integrate them into end-to-end business processes that bring ultimate value to customers. To do this, companies had to integrate their trading partners' functional applications because many companies (suppliers, suppliers' suppliers, etc.) are involved in the value chains that touch the ultimate customer.

AMR Research describes a completely new class of business automation systems, "systems of process" that sit on top of legacy applications that currently, and will continue to, serve as "systems of record." The new focus of business automation is on systems of process, those systems that support a company's own end-to-end business processes.

Carr's article misses the existence of the end-to-end business process and refers instead to functional applications when he speaks of best practices and commoditization. It's just fine if functional applications and off-the-shelf components become best practice, for that's not what counts in creating distinction and strategic advantage-the stuff of distinctive, end-to-end business processes that span the value chain, creating end-to-end processes that matter to customers.

The paradigm shift from applications to business processes in the world of business automation can be confirmed by observing that today every major IT vendor now has business process management (BPM) at the top of its technology stack and that new BPM technology start-ups are emerging, as Oracle Corporation was once a start up during the rise of relational data systems. For some application developers, business process management systems are emerging as the next strategic platform, replacing the database of the last decades. This shift goes all the way down to the theoretical underpinnings of IT, the mathematics.

Relational algebra underpins database management systems and lambda calculus does the same for computer programming. Neither provides the required foundation for dealing with the messy dynamics of business processes. A new math and a new computer science were needed and have indeed been adopted by pioneers in the IT industry. The math is pi-calculus and it underpins the computer science of distributed mobile processes. Both of these developments are not new. They are based on decades of work by people like Cambridge professor Robin Milner, winner of the ACM Turing Award, the Nobel Prize equivalent for computer science. Over the past four or five years, this established body of computer science has been repurposed from scientific uses to commercial business uses and serves as the basis of new information systems capable of handling the dynamics of business processes. Business process management systems can, for the first time in the history of business automation, let companies deal directly with business processes: their discovery, design, deployment, change and optimization.

That is why we are only at the beginning of any sort of IT buildout. We know how to do record keeping with computers, but are only beginning to learn to master business processes, the entity that spans applications, systems, departments and companies; and the entity that will ultimately replace functional applications as we know them today.

Business process pioneer, Michael Hammer, observes that companies know how to do a lot of things that can be understood as processes, such as finding new customers, developing new products and opening new plants. On the other hand, converting general process descriptions into executable processes is difficult for many companies because it is not something in which they have a lot of experience. Improving processes to better serve current customers; using strong processes to enter new markets; expanding processes to provide additional services; taking a process in which you excel and providing it as a service to other companies; adapting processes in which you excel to the creation and delivery of other goods or services; creating new processes to deliver new goods or services-these are activities in which no one has much practical experience. Why? Their cost and complexity were prohibitive in the age of reengineering. This is why mastering the new breakthrough BPM methods and systems changes the calculus for gaining strategic advantage, and those who succeed can look forward to competing during the next fifty years of business and IT.

Grids, Web Services and Computing Utilities The Issue: Grids, Web services and computing utilities don't equate to generic applications-that logic simply doesn't compute

Nascent Web services aren't about a few IT providers creating Orwellian grids and commoditizing the world with massive best-practices utilities. Again, Carr's discussion of commoditization applies to technology vendors slugging it out in a fight to find new ways to deliver computing power. According to Carr, "The arrival of the Internet has accelerated the commoditization of IT by providing a perfect delivery channel for generic applications. More and more companies will fulfill their IT requirements simply by purchasing fee-based 'Web services' from third parties-similar to the way we buy electric power and telecommunications. Most major technology vendors, from Microsoft to IBM, are trying to position themselves as IT utilities, companies that will control the provision of a diverse range of business applications over what is now called, tellingly, 'the grid.' Again, the upshot is ever greater homogenization of IT capabilities, as more and more companies replace customized applications with "generic ones."

Carr's arguments are similar to the invalid tautology: all cats are mammals; all tigers are mammals, therefore all cats are tigers. His argument can be easily construed as: grid computing is an Internet-enabled technology; Web services is an Internet-enabled technology, therefore the grid is Web services, and more and more companies will therefore replace customized applications with generic ones. Huh?

While there was no Microsoft in the 1960s, IBM, GTE and others worked long and hard to create the computer utility where dial-up timesharing provided the same services that Carr describes. Even though, for a short period of time, the new IT capabilities were provided via time sharing to students and academic researchers who otherwise wouldn't have been able to gain mainframe access, most corporations did not use it for their business computing. In fact, IT industry trends point to the exact opposite need. The capacities of computers are increasing while their costs are decreasing. Computers are getting smaller while their capabilities are expanding. There are more specialized and dedicated computing devices, not fewer and more general purpose shared devices.

Who needs a utility via a grid when a multiple-processor, low cost Unix server (or even a cluster of ultra low cost PCs running open source operating systems such as Linux) can deliver comparable performance? Take into account the attendant complexities of service level agreements and multi-user resource management for computer utilities, when compared with the benefits of ownership and control, particularly in light of global cyber-terrorism, and the case has not been made for the grid. In short, Carr's 'grid world'-and consequently his grid-based, ever greater homogenization of IT capabilities-just may never happen. While some forms of utility computing might become successful, they will not be of the kind that will cause companies to replace customized applications with generic ones.

While Web services can be delivered over the grid if it actually does become a commercial success, grid computing isn't predicated on Web services. A quick visit to www.whatis.com can help define the meaning of such technologies, "Grid computing (or the use of a computational grid) is applying the resources of many computers in a network to a single problem at the same time-usually to a scientific or technical problem that requires a great number of computer processing cycles or access to large amounts of data." One of the leading companies that wants to provide grid resources in the commercial marketplace, IBM, provides another clear definition, "Grid computing enables the virtualization of distributed computing and data resources such as processing, network bandwidth and storage capacity to create a single system image, granting users and applications seamless access to vast IT capabilities." Neither site mentions Web services.

The grid is really all about the latest evolution of more familiar technologies: distributed computing; the Internet; peer-to-peer; and an autonomic, self-managing computing model. What has this got to do with Web services and companies replacing customized applications with generic ones? Nothing! If the grid helps companies rid themselves of having to acquire specialized hardware to gain access to computing power, wonderful. Commoditize the heck out of that idea, for with the headaches and complexity out of the way, companies can free up scarce resources and use the grid's commoditized infrastructure, not to diminish the opportunities for competitive advantage, but to increase them. Long live the grid and utility computing, if it succeeds where the 1960s's timesharing didn't. However, buying raw computer power in a manner similar to the way we buy electric power and telecommunications has nothing to do with commoditizing how the utility is used, any more than saying that all who bought IBM mainframes or Intel or Sun servers replaced custom applications with generic ones.

The IT-enabled information society will be driven by visionary companies that see possibilities where others do not and which exploit the low cost, standards-based, plug and play, interconnect-able, networked commodity components to build infrastructures never even predicted. Indeed, Carr admits this possibility, "This war for scale, combined with the continuing transformation of IT into a commodity, will lead to the further consolidation of many sectors of the IT industry. The winners will do very well; the losers will be gone." Here Carr is referring, once again, to the IT industry itself, but today, more than ever, IT does not exist in a vacuum, isolated from other economic activity.

Take just one example, Visa International and its network, (perhaps we should say "grid") which it based on a precise definition of its business practices and processes and which it enabled by IT. Visa now links in excess of 20,000 financial institutions, 14 million merchants, and 600 million consumers in 220 countries. It is reasonable to expect that it will not be the IT industry that creates the next financial services innovation that will transform our lives. We can, however, be equally sure that IT will matter and must play a major role in making tomorrow's innovations happen.

Howard Smith is CTO (Europe) of Computer Sciences Corporation and co-chair of the Business Process Management Initiative. Peter Fingar is an executive partner with the Greystone Group. Smith and Fingar are the co-authors of IT Doesn't Matter--Business Processes Do: A Critical Analysis of Nicholas Carr's I.T. Article in the Harvard Business Review and the landmark book, Business Process Management-The Third Wave. Preview the books at www.bpm3.com/hbr. Copyright Meghan-Kiffer Press, 2003.

Editorial standards