Ted MacNeil writes (>) in reply to me (>>):
>>1. System z (hardware) revenue was up 32 percent
>>for the quarter (year over year); capacity ("MIPS")
>>grew 34 percent. That means System z gained
>>marketshare again.
>You CANNOT say that, as a completely accurate statement!
>You need to know how many new customers those MIPS were
>sold to. If the increased capacity went to existing
>companies, then there is NO increase in share.Actually IBM's CFO said that (gained marketshare), so I'm just repeating what he said. Now, presumably the CFO looks at revenue, what Wall Street, IT market analysts, and other CFOs look at, too. I can understand why. :-) So one would assume he means revenue share. That is, you take the total server market, count it in dollars (sales), and allocate shares among various vendors and types of servers. There's also unit marketshare, that is you take the total number of physical machines and then allocate share percentages to each vendor and by each type, ignoring the price for each machine. Unit marketshare is fairly easy to count, but I think it's becoming increasingly less meaningful with virtualization (if it were ever terribly meaningful). And you can have big debates about how to count: for example, do you count blade racks, blade chassis, blade cards, blade CPU sockets, or blade cores? But I don't think you're talking about unit marketshare either. What you're talking about I think is percentage of companies (and governments) running (and/or buying) IBM mainframes. That's something like "market penetration" or a certain form of "installed base." One big reason an IBM-MAIN reader might care about that metric is it might be a proxy for professional employment stability (whether you might switch employers more or less frequently). There's very little information in the earnings report on that score, but you can logically conclude two things: 1. It is significantly more difficult to increase revenue at such high rates if at the same time the percentage of companies/governments running your product decreases or even stays flat. That's just simple mathematics, just like 34 percent more MIPS at 32 percent higher revenues means a lower average price per MIPS. 2. There was at least one new mainframe customer in 2Q2008: a financial institution in Vietnam. In fact, that was a brand new *country* for the mainframe, and thus it easily rose to the level that a CFO would mention. I happen to know that wasn't the only new mainframe customer last quarter, but that wasn't in the earnings report (individually; of course the revenue is counted). Market penetration has measurement problems also. For example, if you're doing well in an industry that is busy merging, reducing the number of separate companies in that industry, did your market penetration decline? Most rational people would say no, but the mathematics (uncorrected) might say yes. And there are other ways to measure market penetration besides number of companies. For example, you might measure it in percentage of banking transactions (in the banking industry) or some other measurement of business activity. When you get those data, they're really interesting. But it's often tough to figure out that information. - - - - - Timothy Sipples IBM Consulting Enterprise Software Architect Specializing in Software Architectures Related to System z Based in Tokyo, Serving IBM Japan and IBM Asia-Pacific E-Mail: [EMAIL PROTECTED] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
