Data Warehousing

The economy isn't the only reason the data warehouse industry is stumbling.

It's hard to find a technology sector that hasn't been hurt over the past 18 months by the economic slowdown. In the case of data warehousing, however, the recession is only partly to blame.

Even before the economy sagged, data warehousing began to get a bad name because of exorbitant costs, lengthy implementation times and vendors promising far more than they could deliver.

Although a recent market study by Survey.com indicates that overall warehouse spending is on the rise, it also notes that spending on initial implementations has declined by almost 40 percent since last year. Survey.com's Nancy Stewart suggests that the decrease reflects both economic and market shifts, as “more, smaller-sized solutions are being implemented and more packaged software and cost-effective components are becoming abundant.”

Mart Mayhem

Data marts evolved from the need to implement smaller, highly targeted data warehouses quickly. For many, marts were an inexpensive entry into data warehousing.

Steps to Data Mart Success:

  • Adequately staff your data mart projects with a separate team from the one driving your warehouse.
  • Limit the number of data sources to no more than three.
  • You must define a set of policies to prevent data mart proliferation.
  • Be sensitive to business areas with the most urgent need for a tactical solution.

Source: Gartner

Packaged Systems Gain Popularity
The tighter economic conditions have helped fuel growth in packaged data warehouse and data mart offerings. Although packaged data warehouse and data mart software revenues accounted for only $112.6 million in 1999, IDC predicts that this number will rise to $674.5 million by 2004.

The packaged market is not limited to established vendors. For example, in early 2001, the IT services branch of Royal Dutch/Shell spun off a new company, Kalido Ltd.

The company's product—the Kalido Dynamic Information Warehouse—is a packaged data warehouse with a design based on generic data modeling. Whereas traditional warehouses contain one table for each entity, Kalido has only two logical tables: One for business context data, the other for transactional data. It claims to offer greater management flexibility, including XML-based metadata management and the ability to store changes in the business context data without having to make any deeper changes to the database design.

Although Kalido has received significant attention already, analysts agree that 2002 will be a key year for Kalido to gain greater recognition and acceptance.

 

Market Movers

Microsoft
June 2001: Microsoft announces that Verizon is successfully using SQL Server 2000 to power a 9TB customer billing application across four data centers, serving more than 15,000 users.

Our take: Microsoft wants to take midrange marketshare from IBM and Oracle. Proving SQL Server's performance and scalability on larger systems is key to making this happen.

SAS
December 2001: SAS releases SAS/ACCESS Interface to Oracle.

September 2001: SAS joins Microsoft Data Warehousing Alliance, announces support for SQL Server 2000 Analysis Services.

May 2001: SAS releases SAS/ACCESS Interface to PeopleSoft.

Our take: SAS is striving for credibility as a multiplatform, best-of-breed environment, making such alliances critical.

Oracle
November 2001: Oracle releases Warehouse Builder 3i, a wizard-driven tool for accelerating warehouse design and implementation.

Our take: Microsoft is edging into Oracle's midrange marketshare with SQL Server. An inexpensive, wizard-driven tool is Oracle's best defense.

Teradata (division of NCR)
December 2001: Teradata launches Data Mart Consolidation program for integrating departmental data marts.

Our take: Teradata is the Hummer of data warehouses, carrying with it the unfortunate reputation of being too powerful and costly for most normal uses. NCR wants to change this, which is what makes the Data Mart Consolidation program significant. NCR is especially targeting companies that are grappling with disparate Oracle- and IBM-based data marts.

Warehouses Meet CRM
As more complex demands are placed on CRM systems, expect to see more frequent pairings of CRM and data warehouse technologies. The CRM data warehousing software and services market generated less than $4.2 billion in 1999, and IDC forecasts that these revenues will grow to over $20 billion by 2004.

PeopleSoft already offers the Enterprise Warehouse data warehousing solution, although it's essentially for PeopleSoft customers who have committed to the company's Enterprise Performance Management (EPM) suite. Meanwhile, in November 2001, Information Builders released WebFOCUS i-vision Analytical Suite for IBM eServer Systems, an analytical solution for SAP Financial, HR and Sales and Distribution systems.

Average Annual Budget for
Data Warehouse/Business Intelligence
2001 2004
$430,000
$800,000+
Average Data Warehouse Volume
1999 2002
393GB
+2TB
Growth in Data Warehouse Users*
1999 2003
626
2,718
*Average number of users within an organization who access a warehouse.
Worldwide Data Warehousing and Business Intelligence Market Revenue
2001 2005
$60 Billion
$150 Billion

Middle Ground: Logical Data Marts
Not physically distinct from the data warehouse, logical data marts exist only as structures defined in the data dictionary. They require no data replication or synchronization and have the potential to reduce management overhead significantly. Teradata is making strong advances in this arena.

Fundamental Financials: Last 12 Months
Vendor Symbol Sales Income Net Profit Margin Return on Equity Debt/Equity Ratio Revenue per Share Earnings per Share
IBM
IBM
$85.87 billion
$7.72 billion
9%
35.1%
0.80
$49.95
$4.35
Microsoft
MSFT
$26.78 billion
$6.08 billion
22.7%
12.5%
0.00
$4.97
$1.09
NCR
NCR
$6.11 billion
$240 million
3.9%
12.4%
0.10
$62.79
$2.42
Oracle
ORCL
$10.54 billion
$2.5 billion
23.7%
43.8%
0.05
$1.92
$0.44
Sybase
SY
$950 million
$3.8 million
0.4%
0.5%
0.00
$9.62
$0.05
Source: Zacks Investment Research

Earnings Growth Predictions
Vendor FY 2002 FY 2003 Next 5 Years
Oracle
-0.4%
13.5%
20.0%
Sybase
11.2%
n/a
18.8%
NCR
30.1%
n/a
14.5%
IBM
-2.1%
10.8%
12.4%
Microsoft
0.6%
13.6%
16.6%
Source: Zacks Investment Research