This is the first of a two-parter on the information landscape out there in the real world, driven by some strategy thinking we've been doing about our product and how to sell it. In this part, I survey the “Business Intelligence” landscape (it's bad). In Part 2, the question is: how to get people to try new technology in tough times? (Warning: kind of corporate, but I think a lot of people are facing the same issues.)
Business Intelligence · This is a class of enterprise software product usually abbreviated “BI.” Some of the big-name vendors include (in alphabetical order, to avoid hurting feelings): Business Objects (interestingly, the only big software outfit I know of that's based in France), Cognos, Crystal Decisions (based right here in Vancouver), MicroStrategy, and SAS (said to be the largest private technology company).
Nobody Knows What's Going On · If you spend much time out there with the Global 2000, you learn that most of them don't think they're doing a good enough job of managing information. They have databases and portals and search engines and ERP systems and data mines and so on and so on, and yet you'll find big service providers who can't work out which customers, and which branches of which customers, are profitable. You'll find manufacturers who can't (qualitatively) tell you where their built-but-unsold inventory is sitting and why it's over (or under) target. You'll find labor-costing systems that aren't correlated with maintenance-tracking systems that aren't consistent with inventory-control systems. You won't hear many kind words spoken about the “IT” group.
(If you're one of the people we've been talking to and you think I'm pointing a spotlight at you: nope, everyone's hurting, it's bad out there).
You don't have to be a genius to see why this is, and it's not because the companies employ stupid people or are badly managed. It's because using computers to track the real-world ebb-and-flow of business is hard, and if you want to distribute accountability and initiative down your org chart (which you probably do) it's just going to get harder and harder to keep on top of where everything stands.
To make things harder, the last few decades have seen furious corporate merger-and-acquisition activity, so most big companies are patched together out of bits and pieces of other companies. The executives get their pictures on the front of Fortune magazine, and the IT geeks get to clean up after the party. There may be a few big companies that grew in a smooth, organic way to where they are today, but I've never worked for any.
Collecting vs. Using · The bottom line: every company out there is collecting oceans of data on every aspect of their business (I've written on this before). Nobody ever got fired for deciding to retain the records or generate a report.
But investors aren't handing out rewards for collecting a lot of data. It's just meaningless mountains of vacuous bits unless you're getting some good use out of it; and if what I'm seeing is the norm, a lot of people aren't.
BI Isn't There Yet · Those BI vendors I mentioned are, as a group, actually doing well even in the current crappy economic times. Bearing our story thus far in mind, it's not too hard to see why: there is a palpable hunger among the people who run things to know what's going on. But the problem is a long way from being solved. Here are some of the problems I hear about with BI deployments:
- They're slow. You're talking many months, sometimes over a year, to have them up and running.
- Most people don't use them. In practice, these are tools for sophisticated analysts who hide in back rooms and run reports for management. In fact, way back before anyone ever heard of “BI,” SAS was well-known as a vendor of sophisticated statistical-analysis tools, which are still there at the core of the product.
- Sometimes they don't work. I've seen a Gartner study that alleges that fewer than half of all BI deployments actually pay for themselves.
I think the first and third points are related: I have seen deployments of truly great software go south because the process was too big and too slow and the landscape just changed out from under it. The bottom line is, the longer it takes to deploy anything, the higher the risk of it ending up as a career-limiting move for everyone involved.
Markets Abhor a Vacuum · The marketplace being what it is, when you have a vacuum like I've described it just can't last. Whenever there's something that businesses need to get the job done, someone will figure out how to build it and sell it to them. I think the problem is that the vendors in this space are ignoring the lessons of the PC, the spreadsheet, Unix, the Internet, and the Web.
They're trying to go big, to solve the whole problem. But the technologies that come along and change the world never try to solve the whole problem. The PC, when it arrived, had no memory management or user interface or multitasking, it was primitive indeed. But it did a few things well. A spreadsheet isn't a database, but it did a bunch of low-level basic database operations really well. Unix, when it showed up, didn't have a help system or file versions or records in its filesystem or ISAM, all things you could get on other computer systems. But it did a few things really well.
Are you spotting a pattern here? Like those other winners, the Internet, and the Web that rides on it, hit big 80/20 points and avoided the deadly disease of completism. And they all had something else in common: you could put them to work fast, before the landscape changed in progress.
It's Worth a Try · That's what we're trying to do here at Antarctica. Visual Net doesn't do everything a full-featured BI system does, but it's interactive so anyone can use it, and we can deploy it fast: days or weeks (no kidding). Is that enough to win? We know the need is there. Stay tuned.
