The WSJ says Microsoft actually got a second bite at the apple when it was approached about buying Overture in 2003, a time when Google was still plenty vulnerable. Overture combined advertising with search results much as Google does. But Microsoft CEO Steve Ballmer and founder Bill Gates shot the idea down. Yahoo ended up buying Overture, and it now forms the core of the part of Yahoo that Microsoft has been most interested in buying.

Now, Ballmer and Gates are awfully smart guys. They even had the benefit of having had a tight relationship with IBM that gave them a front-row seat as IBM hurt itself in the personal-computer market in the 1980s by making precisely the kind of mistake Microsoft made on search technology. Among other things, IBM delayed using a breakthrough Intel processor because IBM didn’t want to cut into revenue for minicomputers and mainframes. Compaq had no such qualms. It came out with a PC that used the processor and stole the leadership position in PCs from IBM, which never recovered. IBM lost as much as $1 billion a year on its PC business in the 1990s, before selling it to Lenovo.

If Ballmer and Gates goofed, despite their great track record and their experience with IBM, there must be powerful motivations that make people miss opportunities like that presented by Google. And there are.

Our research found numerous situations where an existing business blinded executives to what was coming.

One problem is that the economics of a potential new business are compared with the economics of the existing business, even though the existing business model may soon be obsolete. Kodak, for instance, delayed by years its entry into digital photography because it didn’t want to give up the cushy 65% gross margins it generated from sales of film, chemicals and paper, in return for maybe a 15% gross margin on products related to digital technology. What Kodak couldn’t get its head around was that those 65% margins were going away, no matter what Kodak did, and it needed to adapt to the new world sooner rather than later.

Another problem has to do with size.  Even when executives accept the need to innovate, it is hard for large corporations to nurture new ideas, no matter how good.  One client referred to this as his “Starbucks” problem.  His annual growth target was equivalent to all of Starbucks, making it hard for him to consider an idea that did not quickly add hundreds of millions of dollars of revenue. Another executive talked about his “oak and acorn” problem. Tasked with finding huge oaks, he couldn’t spend much time on acorns.

Two lessons to consider:  First, if you find yourself evaluating the prospects for an innovative idea, don’t limit yourself to what those responsible for protecting existing businesses think. You also need to figure out what the competition can do to you, because it’s unlikely that you’re going to be able to stem the tide of innovation.  Second, make sure that you save some room in your investment portfolio for radical innovations, killer ideas that just might grow into the next Google.  [For more on how to innovate, check out "Unleashing the Killer App," which Chunka co-wrote and which was published in 1998.]

You don’t want to ruin your existing business by letting someone pull a Page-Brin on you. You also don’t want to miss a chance to generate that kind of innovation yourself.

Flush with cash, Microsoft can afford to offer financing at a time when other sources of credit have just about dried up for many businesses. In fact, the financing will cost Microsoft little. Once Microsoft has paid the huge costs for developing software, producing additional copies costs almost nothing, so even if the vast majority of customers defaulted on their 0% financing Microsoft would still come out ahead.

History does include situations where vendor financing went awry. During the Internet bubble, for instance, communications-equipment makers such as Lucent provided financing liberally to startups that wanted to become customers. Once the bubble burst and the startups failed, the communications companies were left holding billions of dollars of bad loans.

Our research also found instances where companies got themselves in trouble offering credit to retail customers. Spiegel, for instance, thought it was diversifying when it gave customers credit cards so they could buy from Spiegel’s catalog and stores. For a time, Spiegel not only propped up its retail sales but also booked healthy profits from its finance arm. But Spiegel was offering credit to people who couldn’t afford it. Eventually, they defaulted in droves. Spiegel’s retail sales plunged, and the company had to go back and reduce the earnings it had booked on its credit cards. The company filed for bankruptcy protection in 2003.

The Spiegel case is the more relevant to Microsoft. While the minimal production costs limit the downside of lending to finance software purchases, Microsoft could be tempted to provide additional financing for services, related hardware, and so on so it can book those software sales. The services and hardware would require risking real, live cash that Microsoft could lose.

If Microsoft can stay disciplined about the customers who deserve credit, though, then the company may be able to win over customers who were on the fence about whether to buy from Microsoft or from a competitor.

Oddly enough, one of the biggest threats to Microsoft has been “freeware” such as Linux, which has undercut the demand for comparable Microsoft products. This time around, Microsoft may be able to take advantage of the power of “free.”

Microsoft’s experiment purports to show that “Vista is great, it is just that people are prejudiced against it.” Microsoft makes its point by bringing together 140 avowed Vista haters who had actually never used Vista. They just concluded from word-of-mouth and from reading reviews that Microsoft’s latest operating system was slow, buggy, etc. When shown a demo of “Mojave,” which they were told was a new Microsoft operating system but which was really Vista, the survey participants were enthusiastic. Thus, Microsoft’s conclusion that the problem with Vista is one of customer perception rather than the actual capabilities of the product itself.

Dan, who knows a lot about experiment design, points out that Microsoft’s experiment was fatally flawed. He acknowledges that getting people to experience Vista without preconceptions is a good idea. The problem is that the experiment provided little of the actual Vista experience, like installation and actual use without the benefit of a canned demo or an expert guide. This, Dan concludes, “transforms the experiment into a PR move and not a true experiment on the capability of Vista.”

Microsoft’s approach is not unlike the market research done by Iridium, the satellite venture that cost Motorola and its partners $5 billion to build. Iridium went bankrupt less than a year after the service was launched, and its assets were sold for $25 million. In a key piece of market research, Arthur D. Little began a survey with this statement: “There will soon be a new personal telephone service which at a reasonable cost will provide you with the capability to be reached or to place calls anywhere in the world using satellite technology, which is not limited in coverage like a cellular phone. To access the service you would have a small handset that fits in your pocket…” When asked whether they’d like such a device, respondents were quite positive, and Motorola enthusiastically reported the survey results at an investor conference. The problem with the survey question, as phrased, was that it was more marketing than market research. The responses validated the appeal of the concept but did little to validate the actual offering. For that to happen, the respondents would have had to understand that “reasonable cost” meant $3,000 per handset plus $3 per minute plus monthly fees, that “anywhere” meant anywhere with a line-of-sight view of rapidly orbiting satellites, and that “fits in your pocket” meant you needed a pocket that would hold a brick.

Motorola and other Iridium investors went on to lose billions because they believed that their market research was valid. We hope Microsoft will not confuse its marketing campaign with true market research.