The Year 2000 Problem and Zitel
This is a discussion of year 2000 companies and Zitel in particular (which, to me, is the best short).
Y2000 stocks are selling something that becomes obsolete in three years. For most clients, the date when conversions must be complete is actually sooner, since most systems use dates in the future. For example, an insurance company must calculate and publish premiums several years into the future; a merchandiser is taking orders and buying inventory with delivery dates several months out. If the dates in my system range a year into the future, my conversion must be done by 1999. If you factor in a year for testing (which most Y2000 experts do), you are back to 1/1/98.
The standard response to this is that Y2000 business will be parlayed into legacy system maintenance and conversions to client/server for years to come. It is hardly a forgone conclusion, however, that such business will materialize or that a staff which has been feverishly specializing in Y2000 work will be the best qualified for these other tasks. There are many firms already doing such work, and competition will be fierce given the armies of Y2000 consultants eager to remain employed. When competition is fierce, profit margins are low.
The Y2000 problem has tremendous urgency, and that is what makes it reasonable to assume there will be a lot of well-paid work to be done. This kind of urgency is very rare, certainly without precedent in data processing. The collision of such a huge project with a perfectly rigid deadline is like the proverbial irresistable force meeting an immovable object. However, there is no such urgency to the fondly predicted follow-on work. Such legacy system maintenance and conversion has been underway for many years, and will continue, but there is nothing about the year 2000 that will make it accelerate. So there is no reason to assume employment or large salaries for anything like the same number of people.
So the consulting component of the business cannot be relied upon after 2000. The software products side of the business is obviously much bleaker. Some argue that even though software for finding/correcting date problems will be obsolete, the software will still be useful for analyzing, documenting, restructuring, and converting legacy systems, because the Y2000 programs are merely a specific application of methods for analyzing systems and applying rule-based algorithms to their description, restructuring, or translation. This is true. But these people reverse the sequence of history. Excellent systems of this sort have been around for over a decade. The new companies are applying this existing technology to the Y2000 problem, and not vice versa. Such systems are indeed very useful for understanding the workings of a large legacy system, but they are no silver bullet, or these doddering old monsters would have given up the ghost long ago.
It is also worth noting a contradiction in the self-promotion of Y2000 companies. The hottest companies claim an automated solution that eliminates the need for tedious, expensive work by programmers. But these same companies say they will thrive after 2000 as their teams of consultants (who have by then established lots of good connections) help firms maintain and convert their legacy systems.
The great thing about an automated solution is that it's cheap. If you write a program that can read a million lines of COBOL an hour and fix all of the date problems therein, you can run that program on any number of computers, or sell it to any number of clients, and where the use of human programmers incurred the same incremental cost for each additional line of code examined, your program solves the whole problem at once for the one fixed cost of writing it (There is some incremental cost in operating the computers, selling the service, etc. but it is small by comparison). At this point, newsletter writers start doing the same kind of arithmetic that alchemists used to do: multiply a dollar a line by a million lines, by the hours in a year, by the thousand computers I will run it on, and... what will I do with the money left over after I own everything on earth?
It never quite works out. One mistake they make in this case is assuming that the cost benefits accrue to the company selling the product or service. In practice, most of the savings will go to the clients. What one fool can do, another can. If there is such a large market for such software and it is so profitable, others will write similar programs and they will compete with each other, driving the price down until it is not profitable enough to attract anyone else. There are already ten or fifteen hot companies, several of which didn't exist a few months ago, and that number will grow several fold over the next six months. Of course, many companies will still make a good living at it for a few years, but not enough to justify even one tenth of their current market caps.
At Zitel's current price of 30, the implied market cap of MatriDigm is over one billion dollars (Zitel owns about one third of Matridigm). For you to buy or own the stock at this level, you must believe that it will go still higher.
To buy a stock this risky, you must believe that it will at least double. They have only two years to earn that market cap (about 2 billion. And remember, because Zitel is a minority owner, one can make no assumptions about how much profit will flow through to them). To do this they must have after tax profits of 80 million per month. Assuming a profit margin of 25%, this requires sales of 320 million per month. So far, Matridigm has announced a single contract (of $5 million, with Nevada, their beta site, with whom they have a special relationship. The press release was extremely vague with nothing about scope, schedule, etc. ).
Because there has been only one small contract and no opportunity for external scrutiny, it would be extremely rash to assume they even have a product that works. But let's say they really do have a good product. They still have to sell it. How long do you think it will take to build an organization that can sell 320 million of services a month? How long do you think it will take to sell each contract? I don't know of any company that has gone from zero to 320 million/month in sales in less than a year. Selling anything with a price tag in the millions takes months or years. So by even the most optimistic assumptions, the 320 million figure is way too low.
Some will say the urgency of the situation makes this a very special case, and that sales will be much faster and easier than could be expected in any other business. That might be true if it were obvious to all that they had a product which does what they claim. But no one can even get a look at this product! Tremendous secrecy has surrounded it. Compare their "sales" approach to the currently common industry paradigm of making demos freely available on the web. Obviously, they know their window of opportunity is limited and they must sell as much as possible quickly; but they have done nothing that would quickly publicize their product and establish its efficacy.
The manager of our Year 2000 conversion project (we are a large NY investment house) has been trying to get information from Matridigm for some time. He can't even get them on the phone! (Try calling them and getting information or arranging for a code conversion: 1-888-YRS-2001)
One can go on and on; the story just gets more absurd. For example, Zitel bought its stake in Matridigm for just a few million dollars. So, if you buy the stock at current levels you are disagreeing (by a factor of a couple of hundred) with the way Matridigm assessed its own value quite recently.
None of this discussion is meant to underestimate the year 2000 problem. It is a huge problem, which will have vast ramifications. But this does not mean any of the hot year 2000 stocks will go up in price from here. Most of the year 2000 expenditures will be done in house by large numbers of low-level programmers. The job is tedious and labor-intensive, requiring a large number of programmers, but (luckily) not the finest. At least fifty percent of the work (by most expert estimates) will be devoted to testing, and for large systems the testing will take a year. All Y2000 experts agree that there is no "silver bullet."
There is an interesting stock market play on the year 2000 problem, but it is not a matter of buying year 2000 stocks. Rather, it goes like this: Because so much money will be spent in the next three years on this problem, much less money will be spent on the usual hardware and software purchases (which come out of the same budgets). Therefore, the stocks of hardware and software manufacturers will be severely hurt. These stocks all have high growth rate assumptions built in, and orders will actually contract. It has been argued that this is why Microsoft and other large vendors have been downplaying and neglecting the year 2000 problem. Some of the more dire forecasts predict a bearish effect on the entire market: a 1 to 1.5 percent decrease in the GNP attributable to year 2000 glitches. |
