Daily Archives: October 22, 2008

The Sourcing Maniacs 2008 Vendor Tour Part IV: BIQ

Yakko I really don’t know why we’re here. Spend Analysis is so yesterday, all the analysts say so. I think the doctor needs to give himself a check-up for recommending it to us.
Wakko Really?
Yakko Anyone can build a Spend Analysis system, all you have to do is follow the old Frictionless Commerce strategy from 2002, and grab technologies off the shelf. You take an OLAP database…
Wakko Like Microsoft SQL Server or Hyperion?
Yakko Right. Then you slap on a front-end OLAP viewer…
Wakko Like Databeacon or Business Objects?
Yakko Yes, exactly. Then you find someone to family and map your A/P transactions…
Wakko Like Grihasoft or Spend Radar?
Dot Yup. Then, ta da!
Wakko Ta da?
Yakko Ta da!
Dot That’s all there is to it!
Yakko Right!
sound of crickets chirping
Wakko I don’t get it.
Dot What’s not to get? You’re done!
Wakko Why?
Yakko Why not? You can drill around your spend! You can even run some canned reports!
Dot Yippee! Drill, drill, drill! Report, report, report!
Wakko I still don’t get it.
Yakko OK, I’ll explain it slowly. You’ve never had this view of your spending before. You were never sure whether all of your IBM spend was really being counted, because there were lots of different entries for IBM in the Vendor Master. Plus, you weren’t really sure what commodities you were actually buying, because the GL code didn’t tell you, and the vendor alone wasn’t good enough. Now you can see, pretty accurately, exactly what you spent with whomever on equipment, on consulting, on maintenance, and so on. And, you can see who spent it, by cost center. And when. And how it was booked to the GL.
Wakko So what?
Yakko What do you mean, so what?
Wakko So what do I do with the Spend Analysis system once I can see all that?
Yakko You take action! You source some categories! You fix process mistakes!
Wakko No, no, what do I do with the Spend Analysis system next?
Yakko With the Spend Analysis system?
Wakko Yes.
Yakko Next?
Wakko Yes, next.
Yakko Well, you… um…
Dot You refresh your data every month, and…
Yakko and… well, you can see how your initiatives are progressing…
Dot and… um…
Yakko er… well…
sound of crickets chirping
Dot Maybe we better ask BIQ after all!
Yakko We’re here … I think.
Dot Are you sure? All I see is a house on a hill. Where’s the big corporate headquarters? The parking lot? The sign on the door?
Yakko the doctor did say that BIQ was, in Ariba terms, a micro-operation organized as a virtual organization.
Yakko knocks on the door
Eric Strovink  Uh-oh! Here comes trouble. What do you guys want?
Dot Define Spend Analysis!
Yakko In twenty-five words or less!
Wakko And then get us some baloney sandwiches! I’m hungry!
Eric OK, let’s start with the basics. Where’s the data?
Yakko In the accounting systems! In payables!
Eric Well, that’s one place. But what about PxQ data on invoices, by vendor, by commodity? How about cell phone usage? Facilities? Fleet vehicles? Equipment repair records? Contract labor detail? HR salary and benefits data? T&E? How about the revenue side? Sales, leases, loans, mortgage data? Insurance claims? Web clicks? Medical records?
Dot That’s a lot of data.
Eric Yes, it is. That’s why BIQ customers typically build dozens of datasets, not just one “A/P dataset.”
Yakko Wait, they build them themselves?
Eric Of course. If you had to pay a third-party services organization to build 75 datasets, could you afford it?
Dot 75??
Eric That’s at just one company.
Yakko But I thought it was hard to build datasets and map data.
Eric There are plenty of reasons to pay someone to build and map datasets, but you should always have the option to build and map them yourself, and it should be easy.
Dot Do people build datasets frequently?
Eric Yes, and they change them all the time, too. Serious data analysis requires flexibility. You need to be able to build datasets quickly — in minutes — and you need to be able to modify them in real time. You need to change their structure on the fly, map them on the fly, and whip them into whatever shape you need for the analysis that you have in mind. And then you need to do the same thing for the next analysis.
Yakko But how can you make those kinds of changes if the dataset is shared by hundreds of users?
Eric You can’t. That’s why sharing a dataset across hundreds of users that’s supposedly going to be used for ad hoc analysis just doesn’t work.
Yakko So much for my off-the-shelf technology idea.
Dot But where’s the server?
Eric There doesn’t need to be one. BIQ runs perfectly well on your laptop, supporting very large datasets — on an airplane, on a desert island, wherever. Or you can connect to a server. Or both.
Dot What about reports?
Eric BIQ populates your own Excel models with OLAP data, then books them by dimension. It’s a much more powerful paradigm than static reporting. And furthermore…
Wakko interrupting
Wait a minute, I think I get it!
Eric Get what?
Wakko I get what you do with the Spend Analysis system next, after you build an A/P dataset! You build more datasets, lots of them. And you modify the ones you have. You keep learning from your data by looking at more and more of it, in new and different ways.
everyone stares at Wakko for a moment
Yakko I can’t believe it! He actually uttered a coherent thought.
Dot I’m stunned.
Wakko I’m hungry.
Eric And I’m going back to work.
Eric Strovink disappears back into BIQ headquarters.