>>>ill the site providers need enterprise class data sources? <<<
Enterprise class. Good point. I feel that there are going to be a number of sites, in the low millions is my guess, that need some database server. Will most of those be quite small?
I'd like to throw out a few really wild guesses and see what the gurus think about them.
US ballpark guesses:
Say that of the 265 million people in the US, 200 million of us do some transactions every day. Lets say 5 transactions a day, including groceries, gas, comic books, school lunch, bridge tolls, flowers, computer leasing, newpaper, etc.
Lets say, again erring to the upside, that 30% of those transactions could and would be done through the Internet. I think that's absurdly high, but go with it for a minute. That would be 300 million transactions a day.
Take another guesstimate: How many sites with small servers and how many large companies or ISPs providing ecommerce services to many small companies on their large servers. Lately a lot of ISPs are collecting many small accounts into one ecommerce server setup that uses one fairly large DB server. Normally *not* enterprise class though, more like departmental size, if those distinctions still mean anything. What I mean by it is, a $50K installation, not a million dollar installation.
So, what kind of power do you need to do 300 million transactions a day? As straight accounting transactions, including shipping, credit card approval, billing, inventory, AR&AP, really not that much.
You could see maybe hundreds of thousands up to 1 or 2 millions of quite small servers that are capable of handling dozens of complete transactions per minute, enough to make me, for instance, rich beyond my dreams. A typical small business that booked an order a minute would have their shipping and accounting systems completely overwhelmed. So an adequate DB would be the smallest part of their problem at that point. Paradox or something similar is probably plenty for the majority of these.
This price range is Paradox to Informix-NT or SQL Server, from a few hundred to a few thousand, with quite a few being free software in one way or another. Total sales of DBs maybe 200 million plus or minus an order of magnitude? Since you have situations like linux/apache with the free DB in that, it's hard to say.
You might see a few tens of thousands of DBs of medium (10K on up) to large size sold for these transactions. Say the average price tag for those DBs is 30k. That would be another billion dollars in DB sales, plus or minus an order of magnitude. This is a lot more significant.
Really this is way overkill for the transaction volume, but lots of people will need their own.
So far you are talking from a few hundred million anywhere up to 10 billion in DB sales over the course of a few years. OK, if the top range is right, that's a year's earnings for the DB business. Add in the rest of the world, and you can see that their could be a years revenues in there, which would help compensate for the falloff in DB purchases for internal use.
What other sources of income might there be here?
Informix has produced a DB that will spend most of its time serving pages, graphics, audio, VRML, for instance. The load for serving media is going to be far greater than the load on DBs from transactions. Problem: Thats not really how DBs have been paid for in the past. People pay on the basis of logins more than features, though that is changing somewhat. This kind of capability may just be what you need to stay in the game, not something you can charge extra for.
What about contracting?
Whether a DB is a small stand-alone, free, at an ISP, or whatever, somebody has to set it up. There are packages that make that easier, but it's never a gimme. Lets say you have to pay $100 and hour for 100 hours per company in ecommerce (including all the little ones and erring to the low side) to set up 3 million companies in ecommerce with a DB. That's 30 billion dollars. OK, plus or minus an order of magnitude. Still 3 billion, a significant fraction of the max upside you could expect from straight DB sales.
(Hey do you like the way I throw these numbers around? Totally accurate, I'm sure. :)
Anyway, my conclusion: Informix needs to start a department with this focused mission: Find cost-effective ways to help companies large and small set up their database-enabled ecommerce sites. Give service by the hour, sell support agreements, hook up web servers, java clients, and what have you. You could support any DB at all, including the free ones. But while you are in there, don't forget to pitch Informix. (Give away Informix to those with service contracts, or would that be bundling?) And don't forget to feed back everything the consultants learn to the ecommerce tools developers.
A lot of the small busines people doing the small sites will indeed be todays software developers and contractors, not the guy who runs the corner store. Be their friend. The lucky ones will have their dites grow beyond their own programming time and capability. Informix is still the best-liked DB company among developers, Oracle the least warm and fuzzy. Don't squander that advantage.
Don't confuse this with other consulting departments, like Oracle does. Don't confuse the mission with the job of selling Informix DBs. Allow in other products. Because this market is worth many times more than the traditional straight DB sales. Which is being commoditized by Microsoft and Borland, and the rest.
So, wudda ya think?
Chaz |
