Wayne Eckerson: The Data Warehousing Institute
- By Jack Vaughan
Earlier this year, Wayne Eckerson joined The Data Warehousing Institute (TDWI) as vice president
of technology services after a long stint at the Patricia Seybold Group where, among other things, he 'wrote the
book' on Web-enabled, online analytical processing. Managing Editor Jack Vaughan spoke with Eckerson on the eve
of TDWI's Leadership Conference in Orlando, Fla.
Are you surprised at the speed with which the data warehouse community has embraced the Web?
No, I'm not surprised because the Web offers so many benefits with decision support that it is kind of a no-brainer
to begin providing that to a warehouse via the Web. There's probably about a half-dozen great reasons to use the
Web, but the reason most people use it is to deploy software on the client desktop. You don't have to send legions
of I/T people out to install, troubleshoot or upgrade software, and that streamlines the process of deploying software
- it reduces overhead significantly. All the logic and updates can be maintained centrally.
The real emphasis here, the real motivation, comes from the I/T department in some respects because it just makes
their life a lot easier. But there are also benefits to users that are driving this. One [benefit] is because the
Web reduces administration and, therefore, overhead cost, companies can afford to deploy sufficient support capabilities
to many more people than they could in the past. What I've been saying is that the Web efficiently democratizes
information access. And we've taken some surveys that reinforce this. Companies are using the Web to provide access
to people within their companies and even to their supply-chain partners, who never had access to the data warehouse
Hasn't the Web client move impacted vendors' pricing plans?
I'm not sure if the pricing has changed that much. What vendors are giving away is kind of a view, but the up-front
cost of the software probably remains the same. It may be a little more cost-effective at higher volumes just because
client/server software wasn't priced for huge volumes of users. So it's more cost-effective that way.
In data warehousing generally, what's working these days?
Well, one of the big trends we see now is the move to integrate data warehousing capabilities and decision support
capabilities into applications. These are SAP, as well as vertical industry applications. And we're seeing vendors
moving aggressively in this space, either OEMing their tools to application vendors or trying to move up the food
chain themselves and beginning to package either templates or reports that get users 80% of the way to the application
they want to build, or actually building the application themselves. The move toward analytic applications is a
big one. Increasingly, people are deciding that business intelligence tools and data marts should be embedded within
the core applications that drive their company.
So, for I/T groups, is the decision going to be that building a big data warehouse themselves is overkill?
Well, it depends what a company's up to. If it's trying to use this application to get a strategic advantage, it
may be worth it for them to build it themselves, because they're going to build capabilities into the application
that aren't available elsewhere. My suggestion is don't build unless you can't find what you need. But remember
that your company is probably not that much different in its core processes than other companies in your space.
You should only build when building would give you a strategic competitive advantage in the marketplace. It's obviously
best to build on a foundation that someone else can support and update as well. You can get some foundation application
upon which to build something else that's best. It's clear that the industry's moving from a build to a buy. And
there will be many more applications out there that can be purchased that can get companies closer to where they
want to be in the end.
Has Microsoft and its Plato software proved a threat? Do they drive the move to applications?
Oh, absolutely. In terms of Plato, people have been expecting this for almost a year. It's very threatening to
Olap server vendors like Hyperion. These vendors definitely feel they have to move out or up. And they're moving
up from technology to applications.
We're starting to see the consolidation people have been expecting. For example, Informix bought Red Brick.
Well, Red Brick was one of those early, high-flying companies in the data warehousing space. They were pioneering,
at least on the vendor side. And like a lot of these pioneering vendor companies, they grow fairly quickly by virtue
of adoption by visionary user companies or early-adopter type companies. But they failed to convert their early
successes into mainstream product. And it was tough from the outset because they were competing head-on as a relational
database vendor against the biggest software vendors in the industry.
So, they've got huge competition - companies that are 20 times the size with 20 times the marketing budget. Their
early successes woke up the sleeping giants to the reality of the warehousing market -their success spelled their
own demise in many respects. As soon as Oracle and others decided warehousing was for real, they made sure Red
Brick was not going to exist. Inside companies, their choice of a database, is a very political, religious choice.
After the early adopters came and left, very few companies were willing to make a choice on a small company like
The way we see it, sometimes there's not much difference between application integration and data warehousing.
At least it seems you have to integrate a lot of applications to get a data warehouse working. Does that resonate
Well, it's funny you say that because I've been saying that application integration is the flip side of data warehousing
for about four years now, and people have been looking at me pretty strangely. But yeah, they are flip sides of
the same coin. It's different approaches to solving the same problem.
See, the problem is we've allowed our companies to become fragmented. We have implemented systems to support specific
applications in specific departments for specific purposes We build up these islands and then we can't connect
them and integrate them.
One way to solve this fragmentation problem is from a process perspective, and we've addressed the process perspective
by implementing a variety of middleware to eliminate all of these point-to-point interfaces between diverse applications.
The middleware gives you a single, uniform interface between all applications that need to share data.
The other way we try to solve this fragmentation problem from a process point of view is to get rid of all the
disparate applications and replace them with a single application like SAP. People were amazed that it's painful
to implement SAP or PeopleSoft or any of these, but I wasn't - because the problem they're trying to solve is not
implementing the new software package. The problem is trying to reintegrate their companies.
Now companies have tried to reintegrate themselves from a data perspective by implementing data warehouses. People
have come to realize that implementing warehouses is hard, and not just because there are lots of technologies
to implement. It's hard because you're trying to eradicate the sins of the past. You're trying to reintegrate a
fragmented enterprise that's taken years to fragment, and you're trying to put all the pieces back together again
in one fell swoop. You're bound to encounter problems.
Some people are talking about a balanced scorecard that gives you a true view of your company's activities.
It seems to have potential as a "killer app" for data warehousing.
Well, a balanced scorecard is kind of the ultimate business intelligence application. It has to incorporate all
the different technologies that we've been talking about for many years - warehousing, Olap, distributed networking
and middleware - to actually pull all the scores from all the individuals, and then consolidate them. There are
a number of companies who play in that space right now. It's a hot area. The problem is that it's more of an organizational
challenge in many respects to prepare a company to access itself using these kinds of measurements and tools. It's
a big endeavor, and you have to be committed to it.
Describe the role of The Data Warehousing Institute.
We are the leading education research organization dedicating to data warehousing. We've have nearly 4,500 members,
who are data warehousing professionals. Our primary activity is four conferences a year. They brought me on to
beef-up the research side, which we're doing through a variety of publications and reports, as well as a network
of consultants I'm putting together. On the research side, the goal is to provide one-stop shopping for any information
users have on data warehousing, and to - between the shows - capture the dialog, interaction and ideas that get
disseminated at our conferences and package that up in an electronic format for users.
Jack Vaughan is former Editor-at-Large at Application Development Trends magazine.