Building a Better Business, Intelligently
Making use of Business Intelligence to make intelligent use of your data
Now that you’re done spending money on Y2K, you can start spending time and money on things that really benefit your organization. The choices are hard, because there’s so much demanding your attention. Many organizations are devoting resources to Internet projects, but because the Internet is evolving at such a rapid pace, deciding where to spend your money is not so clear-cut.
One area where you can invest money and get an immediate return is Business Intelligence (BI). You might ask, “Isn’t that just a fancy name for a data warehouse?” Well, in a word, yes, but it can be so much more. It’s just what the name implies—making intelligent use of your data.
BI usually refers to the front-end tools that are used to present the data in a data warehouse in a more intelligent manner. If your business is doing well, the good news is you have lots of sales reports, inventory movement reports and planning reports. The bad news, however, is you have lots of sales reports, inventory movement reports and planning reports. Making sense out of the morass of data you already have is what Business Intelligence is all about.
This doesn’t mean just taking your reports and sorting them a different way, or downloading the data into an Excel spreadsheet. It means coming up with new ways to look at the data, and new ways of looking at the relationships between data. Some questions you might be looking to answer with BI are: Why is this item selling in one region and not in another? What color is doing best and why? When someone buys one item, what other items are he or she likely to buy at the same time?
In Business Intelligence circles there’s an urban legend about the retailer that found out that there was a high coincidence between the sale of diapers and beer. Apparently, when men went to a store to buy diapers, they also came home with beer. This knowledge prompted the retailer to rearrange store displays to put the beer near the diapers. How did they find this gem of information? By using a BI tool, that’s how!
IBM and other third-party vendors provide a plethora of tools that can do more with data than ever thought reasonable before. With OS/400 V4R3, support for vector indexes was added that increased the speed of accessing data in a data warehouse. The basis for this new breed of BI tools is multidimensional databases, or views of data. Vector indexes assist in improving the data access performance of tables that are formed in this new view.
|Apparently, when men went to a store to buy diapers, they also came home with beer. This knowledge prompted the retailer to rearrange store displays to put the beer near the diapers. How did they find this gem of information? By using a BI tool, that’s how!|
Today, such tools exist on the AS/400 that help perform these tasks. The problem is that you have to think differently. And because for many of us thinking at all hurts, doing it differently is unthinkable!
So how do you get started in this new world of BI? Well, if you’re not yet experienced in BI, the first thing to do is get some help from IBM or another vendor who understands Business Intelligence. It will cost you money, but it will pay great dividends in the delivery and usability of a system. If you don’t understand how to model a data warehouse, or if you think that you’ll be using data replication technology (or if you don’t know what data replication is), you need some assistance.
Not too long ago, we were expected to be experts in technology. But today, we also have to be expert in knowing what we know and what we don’t know, as well as where to get help. Technology is too complicated and changing too fast today for us to even begin to know it all. We can’t emphasize this enough: Get help!
Next, take a look at some of the available tools. Whatever you do, use a tool of some kind. Don’t try to duplicate the functionality of existing BI tools with RPG or Query. You’ll be missing a whole world of innovation and will be limited by the BI knowledge of your programmers. Other people (ISVs) have already put time and energy into research and development, so don’t waste your time trying to develop a custom solution.
The tool is important, but the implementation of the tool (whichever one you pick) is substantially more important. You can take a great tool and install it poorly and be miserable. You can take a mediocre tool and install it well, and be a hero.
Next, you need to define what you want to get out of the project. What is to be accomplished? Here our opinion differs from that of many project planners. If you don’t understand the technology, you will not be able to react to the organizational pressure that new technology implementations engender. Even with a consultant to help, you need to learn how the technology will work within your organization. You may have people who will not adapt to it, and you’ll have to decide how to handle that. On the other hand, you may have people who will eat up these new visions of data faster than you can figure out how to give it to them.
BI applications are resource intensive. If you don’t believe us, take your biggest sales history file, join it to the customer master, item master, pricing master, cost file and journal entry detail, then sort it by descending total margin. We don’t recommend doing this during your order entry busy time! (Do it at lunch when you’re not around.) Our point is that, as you know, this will bring the machine to its knees.
| ||A BI Twelve Step Program 1. Decide if it’s important (if you can’t do this, don’t continue).|
2. Layout a target project (not too specific—you don’t know enough yet).
3. Find some help (this is a maturing industry—there is help out there).
4. Review and select a BI tool set (there are good ones out there but don’t take forever—pick one!)
5. Build and deliver something fast (you may throw it out next year, but get some benefit now, and learn).
6. Layout a maintenance program and data update process.
7. Train your technical staff (and yourself) on how to keep things in sync.
8. Train your user developers.
9. Train your end users.
10. Train everyone again.
11. Deliver the data as quickly as you can (you’ll be asked for more as soon as they figure out what to do with what you’ve given them).
12. Go to step 6 and do it again.
A BI tool tries to help this by organizing the data in ways that optimize data access (not disk storage). You may want to look at picking up a smaller AS/400, just to do this type of reporting. Use your main system to perform the transactions and keep track of what is going on, then replicate the data to a smaller AS/400 that hosts your BI software. This may make a lot of sense today, whereas in the past it might not have.
Next, concentrate on keeping everything in sync. Train the users—or at least a couple of them—to be the BI programmers for the user departments. Ideally these will be application users, not programmers. Virtually all the tools available have excellent interfaces for developing and manipulating the interfaces from a user perspective.
I’m not trying to downplay the importance of keeping the data up to date and clean. That is a technical task and should be handled by the technical staff. However, the reporting is a user task and should be taken care of by the user departments.
Pay attention to what we’ve laid out here and you’ll be well on your way to a successful BI implementation. Don’t enter into a BI project without careful up-front thought and planning. Every hour you spend in up-front planning will be worth ten hours of work on the back end!
John Bussert is president of Swift Technologies (Marengo, Ill), a company specializing in AS/400 and Windows NT software. He may be reached at Jbussert@stecnet.com.
Related Editorial:Business Intelligence—Can You Afford To Ignore It?Business Intelligence a Major Post-Y2K Thrust, Surveys Confirm