[1]

· Information-communications-technology industries experienced double-digit real growth for the third consecutive year, increasing 12.5 percent in 2006, compared with 13.3 percent in 2005. These industries accounted for less than 4 percent of the U.S. economy but more than 14 percent of real GDP growth.[2]

· Ten years ago, broadband penetration in the U.S. was negligible; only 18 percent of U.S. households had any kind of Internet access at all.[3] Five years later, broadband penetration broke 30 percent for the first time.[4] Today, nearly 58 million, or 75 percent, of U.S. households subscribe to broadband, according to the Consumer Electronics Association.[5]

This relentless pace of technology innovation is matched only by the rapid evolution in ways to develop and serve technology to consumers. Here’s how a leading business school publication describes the changes in software delivery:

Traditionally, Microsoft and most other software companies deliver software primarily by licensing “shrink-wrapped” products sold through retail channels or arrangements with hardware vendors. […] But new models of software pricing and distribution are becoming increasingly popular. “Open source” software relies on voluntary programmers to build applications that can be distributed freely. Ad supported software includes web-based applications that are free as well, but they generate revenue through advertisements. Also on the increase: “on-demand” software where customers rent software applications when they need them and pay only for what they use.[6]

In the next section, we examine these evolving competitive threats to the software distribution model used by Microsoft at the time of the antitrust case.

Competitors Derided Then Are Rising Now

“Prediction is very difficult, especially about the future.” – Neils Bohr

By the late 1990s, personal computers were “bulking-up” to carry a heavy load of operating systems and locally installed application software. This was called the “thick-client” model, as distinct from the “thin-client” terminals that had long been used to connect with mainframe and minicomputers.

Microsoft products contributed to the weight gain of PCs by adding graphical user interfaces and an Office suite with ever-growing features. Although other companies were also offering thick-client PCs at the time of the Microsoft antitrust case, Judge Thomas Penfield Jackson carved Apple right out of the relevant market, and laughed at the idea of Linux as a competitive threat to Windows.

Ten years later, Microsoft finds itself in ever-greater competition with Linux-based desktops and Apple computer. Use of these non-Microsoft systems has surged in the years since the Microsoft case began:

· Market research firm IDC reports that paid Linux shipments have grown by 25% each year since 2002.[7]

· Dell, Gateway, and Sony are packaging hardware with Linux operating systems from Red Hat and Ubuntu.

· Market research from Gartner and IDC shows that sales of Linux-based servers are up 42% during the first quarter of 2007 in all of Europe, the Middle East, and Africa.[8]

· Apple now ships 1 of every 6 laptop computers in the U.S., moving up to a third-place tie with Gateway.[9]

At the same time Linux has made inroads on Intel-based computers, Apple has won converts to its proprietary Mac platform, based largely on its well-integrated features. In its TV commercials, Apple claims that Mac users can create websites, use its built-in camera, and edit movies “right out of the box” whereas a Windows user is portrayed as having to download drivers and read manuals before he can use the same features.[10]

While the above indicates much greater competition within thick-client software, the importance of this software model may be on the wane. For millions of Internet users today, the only thick-client software they use is the operating system on their cell phone or PDA.

Moreover, thick-client desktops from Microsoft, Apple, and Linux are feeling competitive pressure from an alternative software model, as discussed in the next section.

From Gateway, the Computer to a Gateway for the Internet

Does it even matter whose operating system I’m running? After all, my PC is just the way I get online.

With the ubiquity of broadband internet connectivity, a “thin-client” model has risen to challenge thick-client desktops from Microsoft, Apple, and Linux. Only this time around, the thin-client isn’t a dumb terminal connected to a mainframe computer. Today’s incarnation of the thin-client is a bare-bones personal computer, PDA, or cell phone that can run an Internet browser and connect to the Internet. As one analyst puts it, “browsers are quickly becoming the most important application in your computer.”[11]

This analysis is supported by the numbers in Harvard Business School professor Marco Iansiti’s report on the case of U.S. v. Microsoft. Iansiti writes that “[t]he competitive dynamic faced by client operating systems have [sic] been impacted over the last five years by the utilization of PCs for internet-centric activities. In 2006, U.S. households spent 31% more time on internet-centric applications than in 2001 while use of client-based applications dropped by 14%.”[12]

Thin-clients are hosting a new wave of Web-based software known as “Web2.0,” which includes “software as a service” or the “on-demand desktop,” where applications are designed as Web-based services to be hosted for use by constantly connected customers.^[13] This desktop trend is the subject of our recent report, The Rise of the On-Demand Desktop.^[14]

The rise of the on-demand desktop will forever change the concept of the desktop market. Centralized Web-based services can instantly deploy new features, integrate across applications stored on the server, and enable collaboration among connected users.

One of the leading proponents of the thin-client, on-demand model is Google, the “superpower of search.” Eric Schmidt, Google’s CEO, sees it this way: “Today’s desktop software will be overtaken by Internet-based services that enable users to choose the document formats, search tools and editing capability that best suit their needs.”[15]

Taken from our report, the table below summarizes the different business models used by four leading desktop competitors in terms of product development, distribution, and pricing.

As seen in the table, many on-demand desktop applications are available at no charge to the user since they are supported by advertising and targeted marketing. To increase their ad revenue, Google offers program interfaces specifically designed to help other applications invoke Google’s search, mapping, desktop and AdWord services.[16]

These new models offer competing ways to package and deliver software features, fundamentally changing the way that consumers access their desktop suite of personal productivity applications. In light of this new reality, the antitrust case’s preoccupation with Microsoft’s thick-client model is out-of-date.

In the next section, we consider how thick- and thin-client models will compete in the next stage of market evolution.

Rocks versus Sunglasses

Try to visualize the thick-client model as hauling around a heavy load of client software, or rocks. You need a lot of disk space and processor power to carry them, and you have to install new rocks every couple of years as updated versions are released. Vendors of desktop rocks like Microsoft and Apple usually spend several years on major releases, and even a core Linux distribution like Debian manages only about one release each year.

Now visualize the thin-client model as pair of sunglasses you use to gaze at the clouds. They’re easy to carry around, and interchangeable since you can switch to another pair of sunglasses and still see the very same clouds.

Obviously, this simple analogy portrays thick-client in a poor light compared to thin-client sunglasses. In reality, the thin- and thick-client models are blurring the distinction, and even Microsoft recognizes the rise of on-demand Internet applications. Ray Ozzie, taking over from Bill Gates as the company’s Chief Software Architect, told analysts that Microsoft would move more of its desktop offerings into “the cloud” of Internet services.[17]

When Microsoft does reach “the cloud,” the company will square-off against the likes of Google. 65% of all Web searches are performed using Google, and the Google toolbar easily integrates with the most popular Web browsers.[18] Moreover, Google is the number one provider of Web-based productivity tools:

Nielsen/NetRatings writes:

[I]n October 2006 the combined unique visitors to Google Docs and Spreadsheets reached 445,762, with an average time spent of 10 minutes per visitor. This accounted for 92% of unique visitors and 95% of time spent among providers of Web-based productivity tools that month.[19]

Thin-client software now works well in comparison to thick-client software. In a review of Google Docs and Spreadsheets, a PC Magazine review reports:

Though Google Docs doesn’t try to match Microsoft Word’s overpowering feature set, it does let you create fully formatted documents complete with pictures, tables, and hyperlinks.[20]

Another review of Google’s on-demand offering in VNUnet says that:

Google Docs & Spreadsheets is a very powerful part of the Google empire. The collaboration tools rival some of the more advanced features of Office and are well worth investigation by any group needing to share documents across a wide area.[21]

The power of the thin-client is not limited to productivity applications. Ten years ago, you bought a box of CDs to install an encyclopedia or mapping software on your PC. Today, all you need is an Internet connection and a handheld device running a browser. A ZDNet review puts it this way: “[T]he user never has to worry about updating to the latest version, since Google’s servers always provide the company’s most up-to-date product.”[22]

Even so, the thick- and thin-client models are destined to co-exist and compete in the near term. Google now offers a download of Sun’s OpenOffice—an extremely thick clone of Microsoft Office—which includes shortcuts to Google’s Web search pages. Even the popular Google Earth program requires a 13 MB download to install Google’s proprietary thick-client interface. And as noted earlier, Microsoft is now investing in thin-client office applications to be hosted on Microsoft Live.

These software technologies and business models ought to be allowed to compete on the merits, without interference from antitrust regulators.

It’s Time to Retire the Microsoft Consent Decree

Over the last ten years, software innovation has been unpredictable and unprecedented, and the results are undoubtedly benefiting consumers. The consent decree governing Microsoft since the antitrust trial shows that it neither predicted nor contributed to these innovations.

Consider the three main themes of the consent decree:

1. Restrictions on Microsoft software distribution deals with computer makers.

2. Required disclosure of program interfaces and protocols to help independent software and hardware vendors interact with Microsoft middleware.

3. Court-ordered review and rules governing how Microsoft integrates new middleware with Windows.

In 2006, after five years of operating under the consent decree, Microsoft voluntarily adopted a set of Windows Principles that it says will promote computer manufacturer and user choice, opportunities for developers, and interoperability for users.[23]

Microsoft General Counsel Brad Smith described Windows Principles as the result of lessons the company has learned through its antitrust troubles:

We’ve learned that people care not only about what we do but about how we do it. So, our goal today is to be principled, transparent and accountable in our design of Windows as we go forward now and in the future. The principles that we’re announcing today assure customers, our industry and regulators alike, that Microsoft is committed to developing Windows in ways that advance innovation for consumers and preserve important and robust opportunities for competition.[24]

One could view these principles as merely a restatement by Microsoft of many of its legal obligations from the antitrust decree. But at least one commentator believes that the Windows Principles are more—an “acknowledgement that Microsoft understands the Web is the number 1 platform now – not the OS.”[25]

Over the last ten years, the operating system has been losing its significance. Harvard’s Iansiti reports that “[t]he number of […] OS-agnostic applications has risen dramatically. […] For enterprises, the number of newly developed OS-agnostic applications grew at 24% CAGR [Compound Annual Growth Rate] from 2000 to 2006 while Windows based applications grew at a 1% CAGR during the same period. The percent of OS-agnostic applications is expected to surpass the percent of OS-specific applications by 2011.”[26]

Viewed this way, the Windows Principles are as much a response to the changing marketplace as they are to antitrust—“promoting competition” not against Microsoft, but by Microsoft. A recent company statement put it this way: “The Windows Principles reflect the degree to which the industry and the company have changed.”[27]

For instance, these Windows Principles commit Microsoft to disclose to the developer community all application programming interfaces within Windows that are used by any other Microsoft product. This means that anything Microsoft’s products can do with Windows, competing products will be able to do as well.

Now that the decree is nearing its expiration date, the U.S. Justice Department and states must avoid tying the hands of one company by extending a consent decree that inhibits integration of features. Nor should they continue to invite competitors to block Microsoft from integrating new features into its operating system and applications.

More importantly, an extension of the U.S. consent decree might continue the requirement for quarterly reviews by the Circuit Court. Those reviews have been turned into opportunities for large competitors to block Microsoft’s attempts to innovate and integrate to meet consumer demand. Consider these three examples of how the overhang of the Microsoft antitrust case is posing obstacles to the integration of features in Windows.

Anti-Virus Security Software

Security features have become critical to operating systems and application software. In particular, firewall protection against malicious intrusions has become a basic necessity as users spend more time connected to the Internet. Operating systems now routinely include firewall security software as standard equipment.

But when Microsoft announced its intent to include anti-virus software with Windows Vista, security software firms protested, citing antitrust concerns. Microsoft’s Windows Live OneCare, a service that combines firewall, antivirus and backup capabilities, was called anticompetitive by Symantec, one of the largest makers of software security products. “I don’t want to say it was monopolistic, but it looked that way to some of us” said Symantec CEO John Thompson.[28]

Portable Document Formats

Applications that allow users to save files in different document formats extend work product usefulness. According to Microsoft, a save-as-PDF feature was one of the features most requested by their customers.[29] Adobe, however, objected to Microsoft’s plans to include a save-as-PDF feature in Office 2007, claiming antitrust and predatory pricing violations. Adobe asserted that PDF export technology was a separate product which, when tied to Office 2007, undercut Adobe’s ability to charge for its own plug-in. Microsoft relented, and the save-as-PDF feature is now available only as a separate download for Office 2007 users.

Desktop Search

Web-based search tools are good for finding information on the desktop too. Yahoo, Google and Microsoft have popular search products that reside on the desktop. But Google is now objecting to the way Microsoft integrated desktop and Internet search features into Vista, its new version of Windows. Six months after Windows Vista was released to consumers and businesses, Google filed an antitrust complaint with the U.S. Justice Department, asking that Microsoft hold back a critical Vista update until their desktop search complaint is addressed.[30]

These kinds of objections and complaints have their desired effect: delay of new releases from Microsoft. Antitrust regulators should ensure that their legal scrutiny promotes competition, and not just complaints from competitors. Yet, quarterly court reviews of consent decree compliance are providing a forum for Microsoft’s largest competitors to dog the company and delay release of innovations.

Moreover, delays in significant software updates are disruptive to independent software vendors and potentially dangerous to consumers and users who are vulnerable to security threats.

Conclusion

Innovation and competition—driven by applications that harness the power of the latest technologies—continue to transform the landscape of the software industry. The on-demand model has risen to challenge thick-client desktops from Microsoft, Apple, and Linux.

Thin-client “cloud” software was not a factor in the late 1990s as it is now in 2007, and its largest proponent, Google, was not even incorporated the day that the U.S. Justice Department filed its antitrust suit against Microsoft. Almost a decade later, Google is arguably Microsoft’s strongest competitor.

Despite the new software environment, government antitrust regulators are targeting the industry for ever-greater scrutiny. Continued oversight of the consent decree in the U.S., the recent ruling in Korea, and the ongoing willingness of EU competition regulators to bring new cases are all signs of regulatory overkill.

Antitrust regulators should avoid overly interventionist policies in today’s on-demand technology markets. In the case of Microsoft, Justice Department supervision of the terms of the consent decree will end on November 12, 2007. We don’t need continued oversight going forward—we have Google and other market forces minding Microsoft well enough.