Computer businesses can only do it once according to what Randall Stross calls the “Single Era Conjecture” in an insightful piece in yesterdays’s NYTimes Read Article. It seems right. IBM didn’t make the transition from mainframes to PCs despite its awesome power. And as Clay Christensen has reminded us in his several books, including The Innovator’s Dilemma, there are lots of successful companies that couldn’t anticipate or effectively respond to innovations that made their core technologies irrelevant.

The poster child for this proposition is the communication behemoth, Western Union, which blew off Alexander Graham Bell when he came a ringin’ with his new patented phone technology. It isn’t hard to see the problem. Most companies do some things really well. But that doesn’t mean they can do other things really well – no more than Coco Crisp can take on Roger Federer. Most companies should stick to their knitting.

Moreover, most successful entrepreneurs build companies that can execute on their idea, and don’t have the luxury to develop a flexible innovative culture that can generate more hits. Indeed, many successful companies, and their employees, get so wedded to their technologies that they just cannot imagine anything new, like the Cobol programmers I’ve known who thought PCs were a passing and inefficient fad. All that said I don’t accept the proposition that IBM couldn’t have been the next Microsoft, that Microsoft couldn’t have been the next Google, or still can’t, or that Google will necessarily last but an era.

One of my favorite examples of a company that keeps getting hits is American Express. It was a leading pony express company in the 19th century, had another hit with the travelers cheque in the early 20th century, and then had another hit with its charge card in the late 1950s. It is interesting for another reason—it captured the payment card industry in the 1960s and 1970s, and is one of the largest card issuers even today, despite entering 8 years after Diners Club created the industry and dominated it riding on strong network effects.

So, is it impossible for Microsoft to take on the new Internet companies? Tough, for sure, but impossible no. Likewise, Google could end up like IBM in a decade, or it could keep on truckin for the next 150 years like Amex. So my more complicated series of conjectures are: (1) Most companies can only do a narrow set of things well and can’t make the transition to anything new. (2) Companies in the computer business are like companies in other businesses. (3) A tiny percent of extremely successful companies, just like a tiny percent of musicians who get one hit, will go on to get another hit. And, Apple should have taught us that you shouldn’t necessarily count anyone out in this business until they are really out.

So, do you buy the Single Era Conjecture?