Publication Date: July 15, 1996
Industry Brainstorm: What's Hot in Data
Warehousing?
Boy, it's hot out there. The data warehousing
market, now valued at about $3.5 billion, is expected
to more than double to $8 billion
before the end of the century. With this in mind, we
asked a panel of experts: "What are the trends
to watch in data warehousing?" Their answers
varied, but the general consensus is that the
Internet and intranets present enormous opportunities
in this field. Other issues - such as data mining and
metadata strategies - emerged, too. The experts'
views appear in summary form below.
Larry English
President and Principal
Information Impact International, Inc.
"There are three major trends emerging that I
see important to data warehousing," says
English: "Data mining, data quality and
convergence of Internet/intranet technology with data
warehouse. Data mining is where you are going to get
major value from your warehouse to discover answers
to 'unasked questions.' Data quality of your
warehoused data will limit or enhance the value and
use of the data warehouse. Inaccurate or missing data
may cause both a loss of credibility and disuse of
the data warehouse, as well as inaccurate pattern or
trend discovery by data mining tools. What is
happening with the World Wide Web and the marriage of
warehouse information with intranet and Internet
technology will multiply the data value through
multiplied accessibility." But while those
issues are in the forefront, "the overriding
critical success factors are sound management and
design principles," English says. "One
potential pitfall is looking at technology as a
silver bullet. The technology is an enabling tool
not the solution."
Richard Finkelstein
President
Performance Computing, Inc.
Business is moving toward "doing data
warehousing and data mining on intranets and the
Internet," Finkelstein says. Intranets will
allow managers to use simple Web browsers for
queries, enable information analysts to publicize
their work more easily, and provide opportunities for
collaborative data warehousing among company
departments, he says. "Intranets make data
warehouses far easier to deploy and maintain and
use." As for the public Internet, Finkelstein
believes "data mining information will be used
in more creative ways to give customers information
about products." For example, a financial
institution could ask a customer to provide a
"profile" on-line, then mine its data to
provide that customer with a list of recommended
banking or investment services. The opportunities in
this area are endless.
Warren Thornthwaite
Managing Director
InfoDynamics
"The biggest area right now is the intersection
of data access tools and the Internet," says
Thornthwaite, who sees a move toward Web-initiated
report and retrieval and the development of related
tools. "Current tools provide full ad hoc query
and reporting functionality on the desktop. The
nature of the Internet right now is such that you
cant have as much interaction on the client
side it's basically report browsing and
retrieval.
[So] the next phase will be the
creation of tools distributed through the Internet.
These would include a client piece perhaps a
Java-based applet - that would provide a true,
interactive, GUI-type interface to build a (report)
query." One possible consequence of such a
development: Existing tool vendors would have to
retool their business strategies to compete,
Thornthwaite says.
Scott Semple
President
NewTHINK, Inc.
"The big thing is data warehousing and the
Internet," Semple says. Why? Because of the
tremendous access to information this technology
affords. In Semples model there are three
"generations" of data warehousing:
establishing the warehouse and tinkering with data;
providing universal access beyond the LAN
environment; and pulling information into the
warehouse from outside sources. But the move toward
the Internet, and ubiquitous access, should not be
viewed as a simplistic answer. The underlying
question, Semple says, must be: "How are people
going to use this information to be better knowledge
workers?" The bottom line for business is that
"the only competitive advantage is better
information," he says.
Sid Adelman
President
Sid Adelman & Associates
"Repository and metadata is an important
issue," says Adelman, emphasizing that if a
company does not have adequate data about data,
"you wont be able to build anything but a
simple, trivial data warehouse." At the heart of
the repository issue are several management and
personnel decisions: Where do you put that
repository? Do you use a full-blown repository or
smaller, interconnected repositories? How do you
integrate the information? And do you have the staff
and expertise allocated to maintain this? "As we
get more and more [types of] data into the data
warehouse, we have to better understand the genealogy
and complete characteristics of the data,"
Adelman says.
Aaron Zornes
Executive Vice President
Application Delivery Strategies
META Group
The focus has shifted from justifying the cost of
data warehousing to delivering data warehouses as a
strategic business imperative, Zornes says. "As
data warehouses expand dramatically in the number of
applications, users and size of database, they are
increasingly becoming a strategic business initiative
(versus merely an interesting technology for mass
adopters or a secret weapon of the technology elite)
with emphasis being put on the business
solutions: customer service, database marketing,
supply chain logistics, and quality management,"
Zornes writes in his report on DCIs Data
Warehouse World & Expo, held in June in Santa
Clara, Calif. "Further out, the next high-ground
in data warehousing will require vendors to add
better support for data mining, text/multimedia and
advanced data visualization."
Warren Thornthwaite, Larry English, Richard
Finkelstein, Sid Adelman, Scott Semple and Aaron
Zornes are featured speakers at DCI's Data
Warehouse World.