The theme of this series has been that, in today's computer industry, platforms beat products every time. So, if you want to take over the world, you have to get other people to support your product, use it, and build on it. In other words, you have to get them to "buy in" to your platform.
"Buy in" means business. While some people see computing as a science, an art, or an entertaining hobby, it is actually a trillion dollar business. In the long run, what matters is the number of people plonking dollar bills, pounds, euros or yen in suppliers' hands.
But anyone distressed by the naked commerciality of "buy in" should appreciate that what it replaced was "lock in", and that was worse. To get "buy in", you almost always have to support other people's standards and platforms, whereas with "lock in", the more incompatibilities you can get away with, the better.
The differences are partly historical. In the beginning, the computer industry was vertically integrated: each company tried to do everything, from developing its own chips to running your data centre. IBM did this better than anyone. It not only made the whole kit and caboodle, it installed it, maintained it, and would even lend you the money to buy it. (See Schofield on Saturday, February 1.)
Although this system had many advantages - you knew where the buck stopped - it had some huge disadvantages. The first was lock in. Once you had bought, say, an all-IBM outfit, it could become impossibly expensive to switch to a different supplier. Like an inhabitant of the old Soviet empire, your only way out was to defect, leaving most of your goods and chattels behind.
In the 1980s, the bulk of the IT industry shifted to a horizontal structure where different companies provided different parts of the product stack. Examples included Intel processors, Microsoft operating systems, Novell networking, Oracle databases, and so on. It may be hard to change one layer in the stack - from Microsoft Office to OpenOffice, for example - but it's nothing like as traumatic as changing every layer at once.
Another disadvantage of vertical integration was that it could make life hard for third-party suppliers. For example, suppose you set up a business to supply a product for IBM S/360 mainframes. If you were successful, you could expect IBM to produce an equivalent product. Indeed, this is exactly what many IBM computer buyers would expect IBM to do, and they would welcome it. This seems short-sighted, because it discourages other companies from supporting the range. But it is common. Look, for example, at the warmth with which Apple Macintosh users welcome Apple products that compete with its third-party supporters. Not even retailers are safe. Well, how much effort would you put into selling Macs if you expected Apple to open its own store down the street? A platform, however, absolutely depends on third party support: on "buy in".
One of the things IBM recognised when it launched the IBM PC in 1981 was that one company could not supply everything users wanted. It therefore followed existing market standards, and invited third parties to produce add-ons and peripherals. The PC division even encouraged the growth of independent retailers at the expense of the IBM sales force (which is not to say the rest of the giant IBM organisation necessarily went along with it, but it did have some effect).
In the long term, helping your enemies can be the best way to help yourself, because if a platform is successful, everybody wins. The rising tide lifts all boats. Of course, in today's computer businesses, most companies are much more interdependent than IBM, Digital Equipment, Wang and other vertically-integrated giants used to be, so there are few recent examples of "buy in" versus "lock in". However, it is still possible to lean one way rather than the other, as happened in the browser wars between Netscape and Microsoft.
Netscape, founded in 1994 as Mosaic Communications, wrote a browser as a way of monetizing the NCSA Mosaic browser created by Marc Andreessen and Eric Bina at America's National Centre for Supercomputing Applications in Urbana, Illinois. Its strategy was to own the browser market, so you could only get Navigator (the browser) from Netscape, and you couldn't change it.
Although the browser was, like Mosaic, free, Netscape planned to make money by selling lots of server software to work with it. Also, the number of users obliged to visit Netscape's web portal would create a valuable web property. Netscape could not literally lock users in, but its actions tended to lock verybody else out.
The other part of Netscape's strategy was to be "cross platform" (Unix, Mac, Windows etc), so it didn't follow Windows standards or integrate with other Windows products. This was a fundamental mistake: Netscape was putting the majority of its effort into supporting a tiny minority (less than 10%) of the market. For Microsoft, however, competing was a no-brainer. Microsoft is a platform company, so it set about building a better Windows platform.
In an earlier column, I discussed the concept of "the whole product". This is simply a model that helps you describe how consumers make purchasing decisions (Schofield on Saturday, January 25). It is useful because buying decisions are influenced by things that are not visibly parts of the product. For example, car buyers look at other things besides the time it takes to get from 0-60mph. These can include the price of insurance, whether spare parts are readily available, the image created by advertising, and whether the manufacturer is about to go bust.
All these are part of "the whole product". Companies that want to succeed therefore have to focus on other things besides innovation and technical excellence. They often have to enhance "the whole product" by enlisting the help of third-party suppliers. These are the kinds of thing platform companies do, and Microsoft did them.
To quote what I wrote in the summer of 1997, Microsoft's marketing plan "aimed to satisfy users (it works, you don't have to download it, it's free), internet service providers (easily configurable, has an administration kit), online services like AOL (object-based and can be customised to add proprietary features), web masters (HTML standards, automation of plug-ins via ActiveX, subsidised access to sites like the Wall Street Journal and ESPN), programmers (can use their existing skills and OLE expertise), magazine publishers (free code for cover-mount CDs), PC suppliers (free extra feature with pre-loaded Windows 95) and many more."
All these extras were aimed at creating "buy in" because, as I added way back then: "given roughly equivalent technology, the product that more people want to win will generally win." This is not, I concede, a particularly clean example of "buy in" versus "lock in". However, it is by far the most interesting example available, and it is illuminating.
First, notice that merely giving away and/or bundling Internet Explorer with Windows was not a winning strategy, and failed to dent Netscape's monopoly market share. Users continued to download Navigator rather than use IE1 or IE2.
Second, Microsoft didn't focus on "buy in" until its browser was as good as Navigator and, in some respects, better - which it was around the middle of 1997. Netscape was as good as dead just 18 months later, when it was taken over by AOL.
Third, Microsoft made decisions that favoured the platform even when they were to the detriment of its users. ISPs, for example, loved the IEAK (Internet Explorer Administration Kit), because it let them customize and brand the browser. But users are almost universal in their hatred of NTL and Freeserve and other branded versions of IE, especially after they have switched ISPs.
Fourth, Netscape knew what it was trying to do, which was specifically to create a "whole product". Its strategy came straight out of my favourite Silicon Valley marketing books, Geoffrey Moore's Crossing the Chasm and Inside the Tornado, so you can read it yourself. When I went to see chief executive Jim Barksdale in Mountain View, California, we quoted bits at one another. When I asked him if he was in the bowling alley, he told me what all his pins were.
Netscape's real problem was that it didn't have a clue about Windows or how the Windows business operated. Barksdale had worked at IBM on mainframes, at Federal Express, and at McCaw, a mobile phone company. Andreessen and Bina had worked on the Unix (X Window) version of Mosaic, not the Windows version. Mike Homer, the marketing man, had worked at Go (a failure in pen computing) and Apple. Apparently, none of them could see that their strategy made it impossible for Netscape to compete with Microsoft. They did not fail by accident, but by design (See chapter 9 of High Stakes, No Prisoners by Charles Ferguson for "a strategic analysis of Netscape's failure").
To anyone who understood the concepts I've been discussing in this column, Netscape's demise was therefore no surprise. Ferguson, who dealt with the company, says it died "primarily because it was unspeakably arrogant and stupid": "They were heading into combat without a helmet. I was trying to warn Barksdale that Netscape was about to get shot to pieces, and he wasn't listening."
Of course, the fact that Netscape deliberately chose a losing strategy while Microsoft chose a winning strategy doesn't tell you whether or not the fight was fair. Owning the Windows platform does give Microsoft a lot of power to abuse, and as Al Capone is supposed to have remarked: "You get a lot more with a kind word and a gun than you do with a kind word alone."
