Originally published in Upside

A great deal has been written about Microsoft’s enormous profit margins and how well-positioned the company is for continued profitable growth.  Most of these stories, however, focus on Microsoft’s core business – software, despite the fact that Microsoft’s most profitable products are its hardware products – the mice, keyboards and joysticks that go out under the Microsoft brand name.  Microsoft’s Hardware Group employs only 110 people, but it booked over $350 million last fiscal year.  This means that each hardware microserf is generating an astronomical $3 million in revenue.  That’s not only the highest revenue per employee inside of Microsoft, it’s among the highest in the entire computer industry.  “We’re one of the most profitable divisions in the company,” Beth Featherstone, the Hardware Group’s marketing VP says modestly, “So they’re happy to let us go off and do our own thing.”  

Given this stellar performance, we have a modest proposal for Bill Gates.  Microsoft should become much more of a hardware company, perhaps even going so far as to buy a hardware company such as Apple Computer or NCR.  Such a move could eat up the mountain of cash Microsoft now has languishing in low interest accounts and move the company into lucrative markets that it might have trouble penetrating otherwise.  Such a move could even build Microsoft into a true revenue giant, increase earnings and its stock price, and expand the awareness of the Microsoft brand name.

Absurd, you say?  Isn’t Gates supposed to be the epitome of a software-only strategist?  It’s true that for years, he’s been critical of companies that attempt participate in both the hardware and software markets.  For example, when IBM purchased Lotus, Gates remarked that the deal was “kind of sad,” further stating in PC Week (6/12/95) that “it’s nice for hardware companies to be independent of software companies.”  In any case, wouldn’t a Microsoft-branded PC create a fiasco by alienating Microsoft’s biggest customers – the hardware vendors that would suddenly become competitors?

Perhaps.  But on closer scrutiny, it becomes apparent that there are many ways that such a move could be accomplished, and in fact, there’s already evidence that Microsoft is moving in that direction.

Microsoft tends to downplay the success of its hardware division and carefully avoids any statement that might indicate that the company has an interest in participating in the broader hardware market.  However, the truth is that Microsoft is very much involved in the design, manufacture and marketing of hardware products, and this involvement, which is largely hidden from the press and public, is only likely to grow over time.

Mice Guys

To understand Microsoft’s broader hardware strategy, it’s essential to first understand the company’s current hardware business.

Microsoft entered the hardware business in 1983 when Gates discovered that the hardware vendors weren’t making a mouse that would support early versions of Windows properly.  The Microsoft Mouse was thus originally a purely tactical product, and probably would have been discontinued if it hadn’t proven to be so successful.  “As Windows took off, the mouse business took off as well,” says Featherstone, “When we became a fairly reasonable group from a revenue business, it only made sense to expand and deliver more product.” Today, Microsoft’s Hardware Group produces several varieties of mice, joysticks, an ergonomic keyboard, and even a automated toddler toy – a computerized version of Barney the dinosaur.

When it approaches a new hardware product, Microsoft designs the hardware but it doesn’t actually manufacture the physical product.  Manufacturing is outsourced to other companies, most in the far east, thereby making it unnecessary for Microsoft to undergo expensive tooling and retooling for new products.  Microsoft coordinates the manufacturing and shipping of products, utilizing its channels connection to help make the products successful.

A key element of the success of these products is the Microsoft brand name. Microsoft’s brand name has an enormous value in the marketplace. That became particularly clear when Microsoft market tested the new Barney toy.  “When we asked the moms of toddlers what they thought of Microsoft,” says Featherstone, “the #1 answer was that they knew it would be a quality product.”  Because of this, Microsoft’s hardware products prominently display the company name on the packaging and on the product itself.

As a result of this brand name recognition, Microsoft need spend comparatively little on hardware marketing and advertising.  Instead, its hardware products to be carried along with the wave of Microsoft’s market success.  “The dominance of Windows and Microsoft’s excellent packaging and advertising gives them a real advantage in this market,” says Mark Regberg, Vice President of Massachusetts-based Venture Development Corporation, a market research firm that studies the computer accessories market.  As a result, “they’re clearly the market leader in input devices, says Regberg.

There is, however, a darker side to the Hardware Group’s success, according to Gary Reback, an antitrust and intellectual property lawyer who specializes in legal battles against high tech firms.  “Microsoft creates an unfair competitive advantage by having API’s that they don’t release to anybody outside the corporation.  That’s how they took over the productivity software market.”

For example, Microsoft was recently accused of questionable market practices when it released the 30-page keyboard specification for Windows ’95.  That document outlined a new set of keys and key combinations which would be required for full usage of the operating system.  The new keys were, according to Microsoft, designed to provide users with better access to active programs.  The specification also laid out new trilograms (like Ctrl-Alt-Del) in the new keyboard designs.

The problem with the new scheme is that it made it more difficult for other keyboard manufacturers to compete with Microsoft’s own keyboard product.  Retooling at rival vendor Sejin, for example, cost $3.5 million, a cost that’s difficult to recover when the new 104-key keyboard would only add about $5 to the cost of a new system.  The added expense without a corresponding revenue increase resulted in a shakeout of the keyboard business, with Microsoft moving into the top tier, selling over 1 million keyboard units in the first 15 months.  Microsoft, naturally enough, denied that it was practicing any proprietary hanky-panky.  “The natural keyboard invented the keyboard upgrade and replacement market,” says Featherstone, “Our competitors have benefited from this as well.”

Be that as it may, the keyboard controversy emphasizes the fact that Microsoft sees it’s hardware business as being strategically connected with its software business.  As Featherstone put it, Microsoft is looking for “breakthrough technology that’s a combination of hardware and software, (which is) where MS is well poised to lead.”  The idea that Microsoft should be involved in the union of hardware and software was echoed when Microsoft responded to the keyboard controversy in Computer Shopper Magazine (August 1995) by pointing out that Apple keyboards have had operating-system-specific Apple-logo keys for years.  That’s a telling remark because Apple’s entire business model is based a proprietary hardware-software pair, not on an open systems model.

Such statements raise the spectre that Microsoft might, at some future date, release a Microsoft-branded PC in imitation of Apple’s early Macintosh product.  At that time, the superiority of Macintosh’s software over MS-DOS gave their hardware a competitive edge and premium price.  In fact, back in the pre-Windows era, Apple’s revenue per employee was $409k which is 10% higher than Microsoft’s Microsoft’s revenue per employee in 1995.  Microsoft, however, maintains that it has no such plans.  As Featherstone put it: “We’re not going to compete with our biggest customers.”

Maybe so, but there is one Microsoft product that has a business model that looks suspiciously a proprietary hardware product, and that’s the recently-released Windows CE.

Trojan Horse

When Windows CE was announced at the Fall ’96 Comdex, it looked like just another operating system – a stripped-down palmtop version of Windows ’95.  There was Microsoft, standing along side its long time partners Compaq, Hewlett-Packard, and Compaq (along some new partners like Casio) touting a family of personal productivity products.  But the resemblance of Windows CE to other Windows products is deceptive.

Microsoft’s other operating systems were developed specifically to run on hardware that was designed and created elsewhere.  Thus the standard Wintel box can not only run Windows, but OS/2, Netware, UNIX, and any number of other operating systems.  That’s not true with Windows CE devices.  It was Microsoft – not the hardware vendors – who determined the basic design of a Windows CE machine.

Rather than adapting Windows CE to existing palmtops, Microsoft prototyped the first Windows CE palmtop and issued a specification to the hardware vendors, who were forced to comply with Microsoft’s design.  The result is a device that only runs the Microsoft operating system and which is optimized Microsoft’s Word and Excel applications.  In other words, Windows CE is a proprietary hardware/software pair, with Microsoft controlling most of the design.

Windows CE machines are also being marketed much as if they were Microsoft products.  Microsoft’s own marketing material feature the picture of a generic palmtop, with the Microsoft brand name and logo clearly visible.  And even though Windows CE devices are manufactured by other hardware vendors, they feature the Microsoft brand name and logo.  On the Compaq PC Companion, for example, the Microsoft logo is larger and more prominent than the Compaq name.  And this doesn’t account for the fact that the first thing that appears on every Windows CE screen is the Microsoft brand name and logo.  In short, if this isn’t a Microsoft-branded PC, it’s certainly the next best thing to it.

What’s striking is the similarity between the Windows CE strategy and the strategy of Microsoft’s Hardware Group.  In both cases, the products were designed inside of Microsoft, manufactured outside of Microsoft, and prominently associated with the Microsoft brand image.  And more importantly, as with Microsoft’s mice and keyboards, Microsoft may be getting the lion’s share of the profit margin on every device sold. “Microsoft probably makes more on each Windows CE device than any of the hardware vendors,” says Rob Enderle, a Senior Analyst who’s the “designated Microsoft watcher” at Giga Information Group, a 400-person market research firm based in Cambridge, Massachusetts.

With Windows CE, Microsoft has done something that’s extraordinarily clever.  It’s created what’s essentially a Microsoft-branded mini-PC while outsourcing the low-margin hardware manufacturing to hardware vendors, all the while convincing those vendors that it’s actually doing them a favor. “The hardware vendors wanted to participate,” according to Andrew Seybold, noted industry analyst, “because if Microsoft comes out with something, the chances for success are much higher.”

This new business model is both brilliant and subtle, and it forms a business model under which Microsoft can expand into new, highly profitable segments of the hardware business, without threatening its current position as the world’s dominant software vendor.  “Microsoft has no choice but to pursue a proprietary hardware strategy when it comes to information appliances,” says Jim Garden, Senior Analyst at Technology Business Research, “The company’s traditional, software-only business model doesn’t work as well when hardware manufacturers control the software that goes onto the product.”

Expansion Opportunities

As long as Microsoft avoids antagonizing the system vendors by releasing a Microsoft-branded desktop, the company can participate in any number of newly-developing hardware markets.  Microsoft can pursue these opportunity through either its Hardware Group’s business model or in the same way that it’s pursued the palmtop PC market with Windows CE.

Microsoft has, in fact, been signaling its interest in non-traditional computing for the past few years, mostly in speeches by Bill Gates.  It’s not too difficult to identify the markets where Microsoft intends to play.  Because none of these developing markets involve traditional PCs, they all represent an opportunity for a Microsoft-driven proprietary hardware/software solution:

·         The Network PC.  Microsoft has been talking about the NetPC for about a year. So far, the company appears to be content merely to publish design guides and system requirements, and add extensions to its operating systems to support the “NetPC.”  However, the idea of creating a Microsoft-branded Network PC was discussed, according to Giga Information Group’s Rob Enderle, who’s been privy to a number of strategy briefings at the Redmond campus.  “They had people looking at it on and off, but it never got very close,” according to Enderle.  The discussions were spawned when it looked as if Oracle were going to build its own Network computer hardware, but when Oracle retreated into its current “we’re only a software company” stance, Microsoft also concluded that the NetPC was too similar to a standard PC for Microsoft to change its long-standing position not to be in the hardware market.  

·         The Kiosk PC.  Another possible area of hardware investment for Microsoft could be the kiosk PC, which Gates, as quoted on Microsoft’s web site, positions as a possible successor to the pay phone: “Anywhere you have a pay phone or a bank machine today you'll be able to simply insert a little credit card device called a ‘smart card’ that will be unique to you. As you put that in, it will let you see your messages, send messages, have a video conference, buy tickets, bank – whatever you want to do.”  The reason that this might be attractive to Microsoft is that such a device would be much more like an ATM machine than regular PC, thus making it a market that the company could enter safely.

·         The TV PC.  The fourth possible hardware product would be the TV PC, a device that allows viewers to plan viewing, pull up information about programs, categorize TV channels and types of TV programs and groups, develop personalized schedules, and cruise the Internet and the World Wide Web.  In this case, however, Microsoft appears to be taking a more traditional approach, working with other hardware vendors – including keyboard rival Sejin – on peripherals and components for future TV PC devices.

·         The Game PC. Another distinct possibility – given the Hardware Group’s current focus on toys -- is a game machine along the lines of a Nintendo or Sega.  “I’d be really surprised if we didn’t see a Microsoft game machine sometime in the future,” says Rob Enderle of Giga Information Group, “It became an issue as soon as Apple made a deal with Bandai.  But last time I was on site they hadn’t identified a device or a manufacturer.”

·         The Wallet PC.  This is a tiny, wallet-sized device that’s used primarily for financial transactions.  Bill Gates talked in some detail about this concept of a wallet PC at Fall Comdex:  “Put away your paper money and coins. In the future you will make purchases with electronic currency (tomorrow's cash) or credit, beamed from your Wallet PC to your local merchant's register-or to someone else's Wallet PC. Your account balance and credit line could be updated instantly or out-of- pocket ‘cash’ expenditures could be tracked.”

Gates’s interest in the Wallet PC is particularly interesting when you consider that Microsoft’s 1996 reorganization created an entire division that was responsibility for only one software product: Microsoft Money.  As most Microsoft divisions have responsibility for a portfolio of similar products, this organizational structure clearly signaled an expansion of Microsoft’s financially-oriented products.

Since then, Microsoft has expanded the Microsoft Investor 2.0 online investing service, which allows users to import portfolio data from Microsoft Money to get prices, business news and historical data on investments.  Microsoft also forged a “strategic relationship with Schwab and Fidelity to provide users with access to online trading.

A logical extension to these moves into electronic commerce would be the Wallet PC, which would be Microsoft-branded, either explicitly (like the mice and keyboards) or implicitly (like the Windows CE machines).  In either case, Microsoft would be position itself to collect the lion’s share of the profit, should the sale of the devices generate significant income.  This could be particularly valuable to Microsoft should the company be barred, through government action, from participating as a financial service provider and collecting a charge on each Wallet PC transaction.

Microsoft’s interest in non-traditional platforms goes beyond tiny, wallet-sized devices. Microsoft recently hired Gordon Bell, the designer of Digital’s VAX minicomputer, for its advanced research group. One of the projects with which Bell is associated  involves something that’s been called a “computing megamachine.” While this is a software rather than a hardware architecture, it’s interesting to speculate what Microsoft might do if building such a machine required specialized hardware device.  Remember, Microsoft actually supplied hardware cards in the past to help in the launch of a new operating system.  When Windows for Workgroups was launched, one of Microsoft’s packages included a network card.

Another interesting possibility for future Microsoft hardware expansion is a Microsoft-branded telephone.  In fact, one of Microsoft’s resellers debuted such a device in Europe that plugged into the PC and used regular phone lines and the TAPI features in Windows.  As with all the other non-traditional PCs and peripherals, Microsoft’s participation in the market may be temporary and tactical –as it was in with the hardware cards in Windows For Workgroups, or it may become the seed of an entirely new hardware business, as was the case with Microsoft Mouse.  Only Microsoft’s top management knows for certain.

An Embarrassment of Riches

Simple economics might also be driving Microsoft into the computer hardware business.

Microsoft currently has over $9 billion in cash reserves, which is getting an average of 5.8% in various treasury instruments.  That money represents about 70% of Microsoft’s total assets, and the low rate of return could become a drag on the company’s overall financial performance.  Microsoft could deal with this situation by issuing dividends, but “that would have major tax consequences for Bill Gates,” points out Dave Mack, President of Technology Business Research.  Like Dave, many analysts believe that Microsoft has little choice but to invest its surplus by acquiring other companies.

The most obvious purchase would be another large software house which would strengthen Microsoft’s software business.  That’s what Microsoft tried to do in 1995, when it attempt to purchase Intuit for $2.3 billion.  However, the Microsoft withdrew its offer under pressure from the U.S. Department of Justice, which clearly feared that Microsoft was becoming too powerful.  Microsoft can probably expect further DOJ interference should it try to acquire any of its other major software competitors.

The Department of Justice routinely reviews most large mergers under authority granted by the Hart-Scott-Rodino Antitrust act.  Microsoft, however, has been able to get around this review by purchasing software houses with a high market value, but with relatively low revenues in the prior year.  For example, when Microsoft purchased Vermeer (to acquire the FrontPage product), it didn’t file with the DOJ because Vermeer’s annual sales were less than $10 million.  The deal, however, was worth an estimated $80 million to $130 million.  Since then, “Microsoft has done other acquisitions of this type,” says anti-trust lawyer Gary Reback.

However, trying to spend $9 billion in this way is like trying to empty a bucket with a teaspoon.  Unless Microsoft starts making some major purchases, its marginally profitable cash reserves will only continue to grow.  Given that the company’s blocked from large software purchases, where can it turn?  Microsoft could invest in something that’s completely outside of the computer industry, like an insurance company, but such an investment would probably be too far outside Microsoft’s core competence for Microsoft to be able to manage it effectively.

Another alternative would be for Microsoft to look outside the traditional computer business.  One example of such an investment is MSNBC, the joint venture between Microsoft and NBC.  But Gates seems a little ambivalent about Microsoft’s foray into media-land.  In an August 1996 analysts’ briefing, he called MSNBC “an anomaly,” and stated that Microsoft was “not becoming a media company in some broad sense.”  NSNBC confuses analysts as well. “I don’t understand the reason behind MSNBC,” says Randall Stross, author of The Microsoft Way, “It involves big money and there doesn’t seem to be a lot of potential return.”

Assuming that Microsoft, to be effective, has to remain inside the computer industry, what else is open for acquisition, seeing that major software houses are taboo?  Microsoft could buy a networking company, but Cisco, Bay, Cabletron and 3Com are booming and aren’t likely to be for up sale any time in the near future. Microsoft could also purchase an IT Services company, but that would only antagonize the other systems house, which Microsoft needs in order to get Windows NT into the data center.  How long would EDS continue to sell NT, once Microsoft was bidding against EDS for the business?

That basically leaves only one available market segment -- hardware. If the acquisition or investment were positioned correctly, there’d likely be no objection from the Department of Justice, which has been perfectly willing to tolerate IBM’s participation in both the hardware and software market segments for decades.   And as long as it could be done without disturbing Microsoft’s current relationships with companies like Compaq and Intel, there’s no reason why Microsoft might not acquire a hardware company.

Providing that Microsoft felt it could run a hardware business profitably, of course.  At first glance, this might seem challenging. Microsoft’s gross margins of 89% dwarf market leader Compaq’s gross margins of around 22%.  However, Compaq’s net margins of around 10% would be better than the 5.8% that Microsoft’s getting on its cash hoard.  In addition, far from being a novice at computer hardware, Microsoft  has a proven track record in the hardware market with the company’s mouse and keyboard business.  In short, there’s little reason to believe that Microsoft couldn’t run a hardware company profitably.

Assuming that such an acquisition is possible, which companies might Microsoft consider?  In fact, there are three companies that might prove attractive, all of which are in a position to be acquired: Apple, Digital and NCR.

Microhard?

The idea of Microsoft buying Apple has a certain weird logic to it.  For one thing, Apple is already heavily into the kind of information appliance that Microsoft sees as the wave of the future.  In addition, the Macintosh represents a significant source of profit for Microsoft.  “They make more money per box on Apple than on Intel,” according to Randall E. Stross, author of The Microsoft Way, which is why Microsoft keeps building Apple software, even though the strategy of the rest of the company is clearly “Windows Everywhere.”  This means that if Apple went under completely, Microsoft would suffer as well.

A more likely candidate may be Digital Equipment Corporation.  “Microsoft has already invested over $100 million in DEC,” points out Rob Enderle of Giga Information Group, “so Microsoft clearly has no problem investing in hardware companies so long as Microsoft doesn’t appear to be a hardware company.”  Curiously, rumors about Microsoft buying DEC were flying around DEC’s headquarters in 1994, according to a former marketing manager, although this might just have been wishful thinking on the part of DEC’s harried employees.

Because most of DEC’s business centers around the Alpha RISC architecture, competitors like Compaq might be willing to view such a purchase as less than a direct threat to their market dominance.  However, while purchasing DEC might not anger Compaq, it might very well annoy an even more important Microsoft partner – Intel.  It’s highly unlikely that Andy Grove would look kindly on Microsoft’s prospective ownership of the Alpha Chip, and Microsoft can’t afford to put additional strain on what, by all accounts, is already a strained relationship.

But the most likely candidate for a Microsoft hardware company acquisition would be NCR. Unlike Apple and Digital, NCR no longer makes personal computers.  NCR’s major source of hardware revenue is its ultra-large, multiple-CPU, mainframes.  These machines are all based upon an Intel chips, so that Andy Grove would be unlikely to complain about Microsoft’s acquiring NCR.  In addition, NCR has other hardware technology that might very well be strategic for Microsoft, such as NCR’s PC-based ATM machines, which could easily be used as a platform for the Kiosk PCs that Gates is so fond of talking about.

And Microsoft has a long term relationship with NCR, which was an early adopter of NT for use in large-scale data center computing. It’s true that NCR does make Intel-based network servers similar to Compaq’s larger models, but Microsoft could easily drop this line of network servers in order to placate Pfieffer. In fact, the only company that might feel threatened by such a purchase would be IBM -- a company whose relationship with Microsoft is distant, at best.

This raises the question whether or not Microsoft really wants to compete with IBM outside of the software market. Three years ago, the answer would clearly have been “no,” but that was when IBM was on the skids.  Today, with IBM looking stronger than it has in years, one wonders whether Gates might not be looking wistfully at IBM’s diverse product portfolio, thinking how much better he could do at the helm of a similarly gigantic firm.

This is, of course, pure speculation.  And Gates’s attitude on hardware companies in general might best be characterized by his showing of a video at the National Press Club in Washington D.C. in November of 1995, which viciously lampooned a unnamed large computer company. 

However, if there’s anything that holds true in this industry, it’s always a mistake to underestimate Microsoft’s flexibility.  There’s simply never been a large company that can change strategies as fast.  Witness the way that Gates refocused the company on the Internet; in a matter of months, its entire direction had changed.  While Gates has, so far, avoided participation in the broader hardware market, there’s no telling when that could change.  At Microsoft, nothing – repeat, nothing – is ever cast in concrete, least of all a strategy that limits the company’s ability to make more money.

Whether or not Microsoft is contemplating the purchase of a hardware firm, the company will continue to compete – and win – inside those segments of the hardware market that it believes it can capture without threatening its current software hegemony. Microsoft, despite all appearances, is a hardware company as well as a software company.  It would be foolish and short-sighted to think that the company has no plans to participate in those market segments that could become significant additions to Microsoft’s profitable bottom line.