There is a puzzle that has for many years intrigued mathematicians and philosophers. It is called the prisoner’s dilemma. Imagine two suspects caught by the police. The cops believe they have both participated in an alleged crime and wish to get one or both to give evidence. As there are no witnesses, this is the only way the authorities will make the charges stick. The cops separate the pair and make each the same offer. If you turn in your accomplice (each prisoner is told) you will go free and your partner will get 10 years in the clink. The basis for this offer is the fact known by the cops and the criminals that the judge in the case has a history of giving reduced sentences and often clemency to guilty criminals who truly show remorse; the judge also believes that co-operation with the police is proof that a criminal regrets his behaviour. Complicating matters is the knowledge on the part of both prisoners that they might each turn the other in. In this case they would not both get off, but would most probably get less than 10 years each, possibly as little as five. And if both prisoners keep their mouths shut, neither will be incarcerated. Suppose you are one of the prisoners in this puzzling situation. What would you do? In a variation on this theme, the problem is compounded by another factor. If both prisoners stay mum, the police have a witness who saw them co-operate on a lesser crime, one that would put them both behind bars for two years. Again, if only one prisoner turns in the other for the greater crime, the cops will drop the lesser charge against the stool pigeon, who will go free. If both turn state’s evidence, the five-year rule applies.

Autobahns

Now what would you do? This kind of puzzle is not merely of interest to intellectuals who study what is called game theory, an area of mathematics advanced by John von Neumann, who also happens to be the father of the stored program computer. There are plenty of examples of this type of puzzle to be found in everyday life. You may think you are a good enough driver to go over the speed limit, but you may also realise that a world in which everyone drove very fast might be a little on the dangerous side. But there are plenty of people who don’t agree; you can find them on the autobahns of Germany, where very fast driving is the norm. You also might find situations in which you can get on public transport without paying. But what would happen if everyone acted that way? And there is a very good example of this puzzle to be found in the computer business. If you get a big discount and nobody else does, you have a real advantage. But if everybody wants (and gets) a discount, you lose your advantage. And if everybody starts getting very large discounts and the vendor loses control of the market, not only do the buyers lose all their advantage, but the vendor loses its corporate shirt. Like the prisoner’s dilemma and its variants, the computer buying problem can have a few complications. For instance, if you are docile and pay full price when a lot of people are getting a break, you suffer inordinately. You are like the prisoner who is honourably circumspect but who gets turned in by an accomplice. When IBM began discounting in earnest, it apparently achieved the immediate goals it had sought. Deals that might have been deferred or even lost got closed. In situations where a competitor was likely to win business that IBM felt it should get, discounts saved the day. At first, customers kept their mouths shut not just because they signed nondisclosure agreements, but because they figured it was in their interest to do so.

By Adam Page

But this phase didn’t last very long. As discounting become not merely prevalent but universal, the game began to turn in favour of the customer and against IBM. The Wall Street Journal once reported a mainframe customer’s remark that the Amdahl coffee cup on his desk might look ordinary but it was in fact worth a million dollars, the size of IBM’s concession on a hotly contested deal. After that story was published, the discoun

ting game, which had long since ceased to be a secret in the computer room, was a factor in the boardroom. Information processing executives could expect to be asked how their deals compared to the deals others were getting. This development forced information processing managers to socialise with their counterparts at other organisations. Gartner Group, among others, provided many such opportunities for its clients and, as getting the right price became a hot strategic matter, it prospered. IBM caught onto this but we don’t think it really understood what was going on. A company that had always been paranoiac about labour unions, it should have noticed that it was suddenly at the mercy of customers’ unions. Consultants are only one source of pricing information the customers have learned to use. There are also leasing companies happy to share information. And there is also the press. And just as IBM has failed accurately to perceive the power of the customers’ unions, it has underestimated the importance of lessors and publishers. If you doubt this admittedly bold assertion, consider the impact of the company’s decision to cease publishing mainframe prices.

Fateful

In the wake of IBM’s fateful decision, the relative handful of customers that had not actively sought all the pricing data available were driven to plug themselves into new information sources. The autodidactic customer, formerly a common problem for IBM’s sales reps, became an absolute fact at every account. Even worse for IBM is the nature of customers’ new affiliations: they are international, not local, subjecting IBM to price comparisons on a global basis. We cannot imagine that this inevitable development is what IBM sought when it hid its mainframe price list. And that is why we are so confident IBM at least at the high levels where policy is made – hasn’t a clue about the nature of the large systems customers on which it depends for a considerable portion of its revenue and, we believe, more than half its gross profit. For the past few years, IBM has played the game (in a mathematical sense) poorly. It has not examined all the options of its customers and the consequences (rewards or penalties) to customers of each option. Neither has IBM examined the impact on its own business of customers’ various choices. It is as if the company in some very deep way has been unable to acknowledge that customers even have choices. It seems IBM believes all it has to do to get its way is to come up with the right marketing plan. Customers have been far sharper than IBM when it comes to game theory or its real world equivalent, common sense. But customers, too, have only recently begun to see the big picture: IBM is not just going through a bad patch, it is in serious, perhaps critical, trouble. A few more years like 1992 and IBM will be plain out of business. Everyone has jumped on the bus without paying and now the whole transport system is going under. It’s too bad John von Neumann isn’t alive any more. IBM used to listen to him.

Copyright (C) 1993 Technology News of America. Published in the April 1993 edition of Infoperspectives International by Technology News Ltd.