Running (eg) $5 million of servers has a bunch of other costs associated with it: physical plant, power, bandwidth, additional hardware for failover, staff, etc. I am not in the business of running server farms, I couldn't tell you what multiple of $5 million gets you the total cost.
But I get your point, and think you are probably right, that _machine_ costs are not the majority of OCLC's expenditures, even including associated costs as above. The majority of OCLC's expenditures (and we could probably look at a public report somewhere instead of guessing) are probably staff, same as most organizations, I'd guess same as LibraryThing.
You pick a particularly unimpressive list of activities for those staff costs (aren't almost all of those people, the lawyers and salesmen and tech support people on the Hitchhiker's Guide to the Galaxy "first ship" that society tricks into going out into oblivion?), but other activities include cataloging and metadata creation, metadata normalization (both manual and automated, which takes staff time to write), other software development (not just worldcat, but the other OCLC software, including software to facillitate ILL, one of OCLC's core services, also that metadata normalization software, software to exchange metadata in both directions with third parties libraries and not (and the staff to negotiate, facilitate, and maintain those relationships); the various APIs they're turning out; R&D, etc), and probably many other productive tasks even I don't know about.
They do provide all sorts of useful stuff. They also charge all sorts of lots of money to lots of libraries. There is also a lot of things I wish they'd innovate faster on, and some at OCLC are certainly trying, but that takes staff resources too, although I also think it takes a willingness to break with the past even if it pisses off your customers that I'm not sure OCLC has (and it would be a risk to try doing that, hoping you do it RIGHT enough to not piss off your customers in the end).
Is what OCLC provides a good value? I really don't know, there is no obvious answer to me. Heck, I don't even know what different libraries pay for OCLC, but even on a collective aggregate level the answer of whether it's a good value is still not obvious to me. Is a count of how many records are in worldcat a good way to decide if it's a good value? No, agreed with Tim. Should OCLC be trying to convince us it is a good value with real info and not propaganda, marketting, threats, guilt, or tugs on heartstrings? Yes. Have they convinced most of the folks I know? No, although most of us are not convinced of the opposite either, we're skeptical but not committed. Are the bulk of decision-makers who decide whether to send OCLC money convinced? I have no idea. Has OCLC better convince them, and possibly change their costs/services such to make it obviously convincing, if they want to survive? Indeed. Do they legitimately want to survive by providing an obviously good value!
, instead of by ensuring monopoly lock-in? I sure hope so, but I'm scared that some OCLC decision-makers may not.
Jonathan
________________________________________
From: Next generation catalogs for libraries [NGC4LIB_at_LISTSERV.ND.EDU] On Behalf Of Tim Spalding [tim_at_LIBRARYTHING.COM]
Sent: Friday, August 06, 2010 11:42 PM
To: NGC4LIB_at_LISTSERV.ND.EDU
Subject: Re: [NGC4LIB] OCLC response to SkyRiver lawsuit
"How many disk reads does it take to catalog an average record? How
quickly do new records get indexed?" Etc.
I don't want to be argumentative, but, simply put, it doesn't matter
anymore. Ingesting, storing and searching tiny text files has become a
trivial task. I'm not saying it's free, but for an organization of
OCLC's scale, it's trivial.
Again, take LibraryThing. We consider ourselves overtaxed parsing MARC
records from the 200+ libraries who regularly upload full dumps. What
does being taxed mean? It means that a virtual server, taking up about
1/3 of a $15,000 box, sometimes gets behind. Boo-hoo for us!
You can assume any multiple you like and the numbers don't work. Let's
imagine, for example, that OCLC needs to do 10,000 times as much
processing as LibraryThing does. (This would imply regular dumps from
2 million member libraries!) Do the math out and you'd need about $5
million dollars worth of servers. Amortize that over a few years, and
it's a percent or two of the OCLC budget. This is an organization that
gains and loses far more from swings in its stock portfolio.
The simple fact is that sending, receiving and processing tiny text
files has become a trivial task. It's been a trivial task for years.
And every year the costs sink still further.
There's no question non-IT tasks, like customer hand-holding,
report-writing, marketing, sales, lawyers, consultants and executives,
still cost money, but to defend OCLC's prices by citing how many
records they store, or how much processing they do, is to be out of
touch with what computers are like today. That argument is dead.
Received on Sat Aug 07 2010 - 00:24:28 EDT