This pattern is a familiar one in markets for digitised products like computer software and music. Those industries have long been dominated by both a winner-take-all dynamic and high turbulence, as each group of dominant innovators is threatened by succeeding waves of innovation. Consider how quickly Google supplanted Yahoo, which supplanted AltaVista and others that created the search engine market from nothing. Or the relative speed with which new recording artists can dominate sales in a category.
Most industries have historically been fairly immune from this kind of Schumpeterian competition. However, our findings show that the internet and enterprise IT are now accelerating competition within traditional industries in the broader US economy. Why? Not because more products are becoming digital, but because more processes are: Just as a digital photo or a web-search algorithm can be endlessly replicated quickly and accurately by copying the underlying bits, a company’s unique business processes can now be propagated with much higher fidelity across the organisation by embedding it in enterprise information technology. As a result, an innovator with a better way of doing things can scale up with unprecedented speed to dominate an industry. In response, a rival can roll out further process innovations throughout its product lines and geographic markets to recapture market share. Winners can win big and fast, but not necessarily for very long.
In this article, we’ll explore why the link between technology and competition has become much stronger and tighter since the mid-1990s, and we’ll clarify the roles that business leaders and enterprise technologies should play in this new environment.
Technology and competition
The mid-1990s marked a clear discontinuity in competitive dynamics and the start of a period of innovation in corporate IT, when the internet and enterprise software applications — like enterprise resource management (ERP), customer relationship management (CRM), and enterprise content management — became practical tools for business. Corporate investments in IT surged during this time — from about US$3500 spent per worker in 1994 to about $8000 in 2005, according the US Bureau of Economic Analysis. At the same time, annual productivity growth in US companies roughly doubled, after plodding along at about 1.4 per cent for nearly 20 years. Much attention has been paid to the connection between productivity growth and the increase in IT investment. But hardly any has been directed to the nature of the link between IT and competitiveness. That’s why two years ago we set out to compare the increase in IT spending with various measures of competition, focusing on three quantifiable indicators: Concentration, turbulence and performance spread.
In a concentrated or winner-take-all industry, just a few companies account for the bulk of the market share. For our study, we focused on the degree to which each industry became more or less concentrated over time. A sector is turbulent if the sales leaders in it are frequently leapfrogging one another in rank order. As well, the performance spread in an industry is large when the leaders and laggards differ greatly on standard performance measures such as return on assets, profit margins and market capitalisation per dollar of revenue — the kinds of numbers that matter a lot to senior managers and investors.
Were there economy-wide changes in these three measures after the mid-1990s, when IT spending accelerated? If so, were the changes more pronounced in industries that were more IT intensive — that is, where IT made up a larger share of all fixed assets within an industry?
In a word, yes.
We analysed industry data from the Bureau of Economic Analysis, as well as from annual company reports, and found that average turbulence within US industries rose sharply starting in the mid-1990s. Furthermore, after declining in previous decades, industry concentration reversed course and began increasing around the same time. Finally, the spread between the highest and lowest performers also increased. These changes coincided with the surge in IT investment and the concurrent productivity rise, suggesting a fundamental change in the underlying economics of competition.
Looking more closely at the data, we found that the changes in dynamics were indeed greatest in those industries that were more IT intensive — for instance, consumer electronics and auto part manufacturers. Further, we considered the role of merger and acquisition activity, globalisation and R&D spending in our analysis of the competitive landscape and found some minor correlations — but none strong enough to override our measures.
One interpretation of our findings might be that IT is, indeed, inducing the intensified competition we’ve documented — but that the change in dynamics is only temporary. According to this argument, the years since the mid-1990s have seen a one-time burst of innovation from IT producers, and it’s simply taking IT-consuming companies a while to absorb them all. Businesses will eventually figure out how to internalise all the new tools, proponents of this theory say, and then all industries will revert to their previous competitive patterns.
While it’s true that the tool kit of corporate IT has expanded a great deal in recent years, we believe that an overabundance of new technologies is not the fundamental driver of the change in dynamics we’ve documented. Instead, our field research suggests that businesses entered a new era of increased competitiveness in the mid-1990s; not because they had so many IT innovations to choose from, but because some of these new technologies enabled improvements to companies’ operating models and then made it possible to replicate those improvements much more widely.
Although modern commercial enterprise systems are relatively recent — SAP’s ERP platform, for example, was —introduced in 1992 — by now, companies in virtually every industry have adopted them. According to one estimate, spending on these complex platforms already accounted for 75 percent of all US corporate IT investment in 2001. More recently, Gartner Group predicted worldwide enterprise software revenue would approach US$190 billion in 2008.
To understand how this profusion of enterprise IT is changing the broader competitive landscape, imagine that a drugstore chain like CVS that fills prescriptions has a number of rivals, most of which also have multiple stores. Before the advent of enterprise IT, a successful innovation by a manager at one store could lead to dominance in that manager’s local market. But because no firm had a monopoly on good managers, other firms might win the competitive battle in other local markets, reflecting the relative talent at these other locations. Sharing and replication of innovations (via analog technologies like corporate memos, procedures manuals and training sessions) would be relatively slow and imperfect, and overall market share would change little from year to year.
With the advent of enterprise IT, however, not just CVS, but its pharmacy competitors have the option to deploy technology to improve their processes. Some may not exercise this option because they don’t believe in the power of IT. Others will try and fail. Some will succeed and effective innovations will spread rapidly.
The firm with the best processes will win in most or all markets. At the same time, competitors will be able to strike back much more quickly: Instead of simply copying the first mover, they will introduce further IT-based innovations, perhaps instituting digitally mediated outsourcing or CRM software that identifies cross-selling and up-selling opportunities. These innovations will also propagate widely, rapidly and accurately because they are embedded in the IT system. Success will prompt these companies to make bolder and more frequent competitive moves, and customers will switch from one company to another in response to them.
As a result performance spread will rise, as the most successful IT exploiters pull away from the pack. Concentration will increase, as the losers fall by the wayside. And yet turbulence will actually intensify, as the remaining rivals use successive IT-enabled operating-model changes to leapfrog one another over time. Thus, despite the shakeout, rivalry in the industry will continue to become more fast-paced, intense, and dynamic than it was prior to the advent of enterprise technology. These are exactly the changes we see reflected in the data. In this Schumpeterian environment, the value of process innovations greatly multiplies. This puts the onus on managers to be strategic about innovating and then propagating new ways of working.
To survive, or better yet thrive, in this more competitive environment, the mantra for any CEO should be, “Deploy, innovate, and propagate”: First, deploy a consistent technology platform. Then separate yourself from the pack by coming up with better ways of working. Finally, use the platform to propagate these business innovations widely and reliably. In this regard, deploying IT serves two distinct roles — as a catalyst for innovative ideas and as an engine for delivering them. Each of the three steps in the mantra presents different and critical management challenges, not least of which have to do with questions of centralisation and autonomy.
The deployment challenge
Since the mid-1990s, the commercial availability of enterprise software packages has added a new item to the list of senior management’s responsibilities: Determining which aspects of their companies’ operating models should be globally (or at least widely) consistent, then using technology to replicate them with high fidelity. Some innovative teams have pounced on the opportunity. Many more, however, have embraced this responsibility only reluctantly, unwilling to tackle two formidable barriers to deployment: fragmentation and autonomy.
Historically, the regional, product and function managers have been given a great deal of leeway to purchase, install and customise IT systems as they see fit. But bitter experience has shown that it’s prohibitively time-consuming and expensive to stitch together a jumble of legacy systems so they can all use common data, and support and enforce standardised processes. Even if a company invests heavily in standardised enterprise software for the entire organisation, it may not remain standard for long, as the software is deployed in ways other than it was originally intended in dozens, or even hundreds, of separate instances. When that happens, it’s almost certain that data, processes, customer interfaces and operating models will become inconsistent — thus defeating the whole competitive purpose of purchasing the package in the first place.
That’s what initially happened at networking giant Cisco. In the mid-1990s, Cisco successfully implemented a single ERP platform throughout the company. Managers were then given the green light to purchase and install as many applications as they wanted, to sit on that platform. Cisco’s IT department helped the various functions, technology groups and product lines throughout the world get their desired programs up and running, without attempting to constrain or second-guess their decisions.
When newly arrived CIO Brad Boston assessed Cisco’s IT environment in 2001, he found that system, data and process fragmentation was an unintended consequence of the company’s enthusiasm for technology. There were, for example, nine different tools for checking the status of a customer order. Each pulled information from different repositories and defined key terms in different ways. The multiple databases and fuzzy terms resulted in the circulation of conflicting order-status reports around the company. Boston’s assessment also revealed that there were more than 50 different customer-survey tools, 15 different business-intelligence applications and more than 200 additional IT projects in progress.
Deployment efforts heighten the tensions — present in every sizable company — between global consistency and local autonomy. As the Cisco example shows, however, this conflict often exists by default rather than by design. Ultimately, the top team’s focused efforts to manage this tension reaped tremendous benefits.
Responding to the CIO’s assessment, senior managers decided to upgrade Cisco’s original ERP system and other key applications to support standardised data and processes. The upgrade was budgeted at US$200 million over three years. Cisco identified several key business processes — market to sell, lead to order, quote to cash, issue to resolution, forecast to build, idea to product and hire to retire — and configured its systems to support the subprocesses involved in each stage. The software updates and the strategy discussions the technology bought about, eventually resulted in greater consistency throughout the organisation and contributed to Cisco’s strong performance over the past few years.
Data analytics drawn from enterprise IT applications, along with collective intelligence and other Web 2.0 technologies, can be important aides not just in propagating ideas but also in generating them. They are certainly no replacement for brilliant insights from a line manager or a eureka moment during a meeting, though they can complement and speed the search for business process innovations.
Part of the attraction of enterprise systems has been the opportunity for management to impose best practices and standardised procedures universally, as CVS did to great advantage, and so eliminate the chaos of inconsistent homegrown practices. There’s really no competitive advantage in having each department develop and use its own idiosyncratic process for inventory control, for instance, especially when best practices already exist.
While an ERP system is an obvious tool for propagation, other technologies are also important — and they show that innovations do not necessarily emanate from headquarters. For instance, Web 2.0 applications can help process changes emerge organically from lower levels in an organisation. Within Cisco, for instance, a community of about 10,000 Macintosh users was dissatisfied with the level of support they were receiving from the company’s central IT group. But instead of complaining, they created a wiki to share ideas about how to use their Macs more effectively. They posted information, files, links and applications that could be edited by any user — tips and tricks that ultimately became huge productivity enhancers. In this case, process innovations flowed through the company to its great benefit without central management direction.
Maximising return on talent
As corporate IT facilitates the implementation and monitoring of processes, the value of simply carrying out rote instructions will fall while the value of inventing better methods will rise. In some cases, this may even lead to a “superstar” effect, as disproportionate rewards accrue to the very best knowledge workers. Human resource policies and corporate culture will need to evolve to support this type of worker. An effective leader and a well-designed organisation will aggressively seek out and identify such individuals and the innovations they generate, as well as develop and reward them appropriately.
The arrival of powerful new information technologies does not render obsolete all previous assumptions and insights about how to do business, but it does open up new opportunities to executives. Our research has led us to three conclusions: First of all, the data shows that IT has sharpened differences among companies instead of reducing them. This reflects the fact that while companies have always varied widely in their ability to select, adopt and exploit innovations, technology has accelerated and amplified these differences.
Second, line executives matter: Highly qualified vendors, consultants and IT departments might be necessary for the successful implementation of enterprise technologies themselves, but the real value comes from the process innovations that can now be delivered on those platforms. Fostering the right innovations and propagating them widely are both executive responsibilities — ones that can’t be delegated.
Finally, the competitive shakeup brought on by IT is not nearly complete, even in the IT-intensive US economy. We expect to see these altered competitive dynamics in other countries as well, as their IT investments grow.
It is not easy for most companies to deploy enterprise IT successfully. The technologies themselves are complicated to configure and test, and changing people’s behaviour and attitudes toward technology is even more challenging. Enterprise IT typically changes many jobs in major ways; this is never an easy sell to either employees or line managers. As the performance spread, concentration and churn increase, management becomes a distinctly less comfortable profession — more unforgiving of mistakes and faster to weed out low performers. Even those executives who are prepared will not necessarily survive the inevitable turbulence. But those who do can expect greater rewards — at least until another player comes along and uses IT to propagate a business innovation that’s even better. Harvard Business Review
Andrew McAfee is an associate professor at Harvard Business School and Erik Brynjolfsson is the director of MIT’s Center for Digital Business.