ANALYSIS: Too Much Data Warehousing?

By John Bussert

Don’t get me wrong, I think that Data Warehousing and Data Marts can bring a lot of benefit to many companies. Actually, any organization can gain benefit from them. I recently was with a client that has invested heavily in this technology. Over the last two years they have hired three people whose sole function is to provide these services to the various departments. They also purchased a number (over $150,000) of software packages (not counting the servers, OS software and other client software) that they needed to present the results.

I went to visit them to see what had happened over the last year since I had last seen them and was excited to see what this Fortune 1000 company had done. What I found was an organization in disarray as far as this technology was concerned. Oh, they were using the software, creating the source data from the main computer, but all they were doing was using the report writer function and creating the same kind of reports that they were getting before.

To their credit, the users were producing there own reports. But all they were doing was producing "green bar" reports just as before. You might ask why this is better than using OS/400 query? My answer is - "Heck I don't know." But it got me to wondering - are we technologied (Yes, I know that is not a real word) out?

These tools are wonderful, and can provide much to an organization, but it is something you have to fully commit to. There are a number of coststhe software, the servers, the client, the people to manage and develop it, the training of the users and the dedicated set of users to analyze and use the results. Even though the cost of much of this has dropped over the last few years, the cost of implementing it has increased with the cost of hiring and retaining qualified technical people.

Every time I sit down with people and show them what can be done with a Data Warehouse tool, they get all excited. First off, it presents information very well. It is slick, graphical, usually fast, can show relationships very well and is just down right neat! But then we start talking about costs, and things go south quickly. We are not talking about a fortune, but we aren’t talking about peanuts either. Costs can usually run around $50,000. When confronted with the costs the response is usually, "Yes we should do that, but right now, our budget will not allow it."

So they just produce the queries and reports that they have been using for years. Is this bad? I don't know. There are a lot of benefits that can come from this kind of technology, but many companies just can not afford (or won't budget for) this kind of technology. Why? I think maybe, that they have a hard time justifying the technology because they can't visualize the real benefit.

Let’s face it. Immediate gratification usually wins. Given the choice between the same “green bar” report they are used to right now, or a much better report later, they will choose the “green bar” report every time. What they do not realize is that once the infrastructure is setup, they can get the improved reports much quicker, with better data, if they setup a data mart of some kind. Still, it is an up front investment, and unless someone else in their industry already has proven the value, they are very hesitant to commit the resources -- both human and monetary.

Worse, even if they implement a data warehouse, unless there is someone very strong leading the charge, they will put in the technology, but use it in a way they have always used it. They will produce "green bar" reports from the data mart rather than using the relational tools that are available. To do that they have to think differently, and we all know how much that hurts!

Oh, well, I guess that is enough rambling and venting. There is an exciting new technology out there that can help organizations, but only if they can see a benefit and will put the time and resources into making it work. If not, well, then do what you are doing and just produce more reports and paper. Oh, and of course wait until your competitor has proven to you that they can grow their market share using a tool like this at your expense. But then you will have a good reference to point to so, you might as well wait.

John Bussert is president of Swift Technologies (Marengo, Ill.), a company specializing in AS/400 and Windows NT software.

Must Read Articles