|
|
|
Ask the expert
Check out some of the past Ask the expert columns and industry white papers on active data warehousing.
|
Operating an intelligent enterprise enabled by an active data warehouse is no longer an option.
|
Supplying intelligence
Supply chain management software improves operations.
|
The .NET result
One way to simplify data access for software developers is the .NET Data Provider for Teradata.
|
Tech support
Understand how to best provision AMP worker tasks.
|
|
|
|
|
|
Active data warehousing: from nice to necessary
Operating an intelligent enterprise enabled by an active data warehouse is no longer an option; it has become a requirement.
by Dr. Richard Hackathorn
In Chattanooga, a voice-response unit at
the call center for UnumProvident clicks
to "active" for the thousandth time that day.
A customer is calling to determine the claim status of their account. Quietly the unit submits a query to the data warehouse to retrieve the current data on this account; in a clear voice, the system gives the customer the desired information within seconds. This is just one example of how active data warehousing shifts the data warehouse from a passive role to an active role.
The three themes
Enabling a company to take action based on
the information in the data warehouse is the key differentiator of active data warehousing. Traditional analysis is essential for long-term assessment of business dynamics; active data warehousing moves the enterprise data warehouse (EDW) forward to minute-by-minute operational support.
In my 2002 paper "Current practices in active data warehousing," I described the practices of active data warehousing through interviews with seven companies who were pioneers in
this area. These interviews revealed several common themes, such as "single view of the business" and the business value of data freshness. In addition, the study identified factors contributing to the success of an active data warehousing operation, such as securing executive support and adopting an open-information policy.
The benefits realized from active data warehousing by the study subjects were numerous: reduced customer support costs, improved
asset management, audited billing practices, terminated unprofitable products, minimized litigation cases and reduced staff requirements.
Earlier this year, I updated that study by conducting interviews with SUPERVALU and UnumProvident to assess the role that active data warehousing now plays in the enterprise architecture. Those interviews revealed that active data warehousing implementations have matured into mission-critical imperatives for these companies, and highlighted the following three key themes:
 | The holistic view of business operations |
| The business value of data latency |
| The multi-temperature model of data usage |
The holistic view of business operations emerges from the concept of a single view
of the business. The data warehouse supports people throughout the company, from top senior executives with dashboards to entry-level marketing analysts to front-line staffers making day-to-day operational decisions. When accepted by the corporate culture,
the data warehouse becomes the hub for
information dissemination.
Merger and acquisition activities at both UnumProvident and SUPERVALU reinforce
the holistic view. "Our greatest challenge was the integration and rationalization of over 34 line-of-business systems, all reflecting differing heritages of the original firms," says Margaret Johnson, manager of database administration for UnumProvident.
Likewise, Rick Collison, director of enterprise data warehousing for SUPERVALU, emphasizes the value of choosing a common system for managing products, pricing, promotion and placement across their various companies. "[The data warehouse] is essential in the
successful integration of our diverse grocery businesses," says Collison.
Time and temperature
Active data warehousing is perceived to be similar to real-time data warehousing, with
a data latency measured in minutes. The simplistic view is that real-time operation equals business value, but that is not necessarily
accurate. As we got into the discussions,
it became apparent that the concept of data latency is complex. Managers continue to struggle with the issue of how lower data latency can be of value to the business.
|
Figure 1
|
|
As the response time of a reaction to a business event increases, the value gained by that action drops.
|
|
|
Figure 2
|
|
The response time consists of the sum of capture latency, analysis latency and decision latency.
|
|
|
Figure 3
|
|
Decreasing the response time shifts the action higher on the value-time curve, increasing the business value derived from the action.
|
|
The traditional reasons for avoiding low data latency within the data warehouse have been cost, system incompatibility and lack of competitive necessity. Today, the technology to support low data latency is steadily becoming more economical and reliable. It has become
a common tool for strengthening the corporate technology infrastructure. Now, the traditional reasons for not considering applications requiring low data latency do not apply.
An increasing number of visible active data warehousing applications illustrate tangible business value. Johnson of UnumProvident is proud that their data warehouse supports the voice-response units for the call center. The benefits were immediate and measurable. Calls handled by the automated system within target time went from 60% to 95%, allowing staffers to be redeployed on more difficult service requests.
Previously, we introduced the concept of the value-time curve, which shows the relationship between the time required to take action in some business situation and the value of that action (see figure 1). A business event happens, then an action is taken. If the business value of taking that action decays rapidly after the event happens, then the benefit gained from expending the same amount of effort and resources drops. The solution is to move higher on the value curve by minimizing the response latency.
Response latency consists of the sum of capture latency, analysis latency and decision latency (see figure 2). One or more latencies must be reduced to effect a quicker action. Technology advances continually reduce capture and analysis latencies. Decision latency, however, is driven primarily by the cultural aspects of how decisions are made in a company. As a result, decision latency will increasingly become the limiting factor in improving response latency.
As quicker actions are taken, we move up
the value-time curve, increasing the value gained (see figure 3). The value-time curve provides a tool to encourage companies to review their corporate decision-making policies. Active data warehousing is enabling companies to evolve key business processes
that are time-critical in their execution.
The final theme is the importance of a "multi-temperature" model of data usage within the active data warehouse. The activity of data processing within an active data warehouse
is considerably higher than in a traditional data warehouse. The frequency of touching certain data is a measure of its "temperature," which is becoming critical to managing data warehouse performance. Commonly accessed ("hot") data must be more easily retrieved than rarely accessed ("cool") data, for example.
It's an important assessment to make since the difficulty of accessing cooler, or even dormant data can add to query costs. Calculating performance metrics using the previous several years of data, say, can be expensive. If that is your business requirement, however, there is no choice but to spend the money.
Lessons learned
A recurring theme that emerged in all of the interviews was the requirement for a strong business sponsorship of the active data warehousing effort. The barriers to effective active data warehousing are often not technological but organizational. Having the right sponsor
in the organization can make all the difference.
Of course, the best way to gain advocates within the organization is by showing them results they can use. Collison of SUPERVALU mentioned the need for alignment of the
active data warehouse to business priorities. Delivering on promises is key. You should strive to outperform your goals, establishing a track record of delivering on your promises.
Another lesson learned involved management of the active data warehouse architecture. It is important to be clear about definitions and change control, says Johnson. She warned about scope creep, in which business users keep introducing new functions. "When they understand the power of the data warehouse, they will start suggesting lots of new functions. Some ideas are wise ones and good use of technology; some are not feasible within project scope. You need to expand into new functions, but do so carefully."
Part and parcel of defining scope is developing a broad long-term plan. "You must deliver something every step of the way," says Collison of SUPERVALU. In addition, he says, "You must be smart about judging the needs of data warehouse users and the timing of those needs. And expect that their needs will evolve and change."
The demands of the global economy require continuous improvement in the ways of doing business. Enterprises must measure what they are doing, discover efficiencies and then set appropriate goals. By providing a holistic view, the active data warehouse has become the information hub that provides the analytics
to support that continuous improvement.
Companies who successfully implement active data warehouses reap the benefits. Lessons learned reinforce the importance of building on incremental successes and slowly educating executives and workers alike.
The concept of data latency continues to dominate the discussions. The technical aspects are emphasized less, however, as the business aspects become the driving force. The focus has shifted to identifying the opportunities for low-latency data to enable quick and meaningful actions by front-line staff. Today, the real challenges of active data warehousing lie in the organizational aspects. To realize the benefits of active data warehousing, companies may have to change the way that they do business.
Despite the challenges and the lessons
still being learned, one fact is indisputable: operating an intelligent enterprise enabled
by an active data warehouse is no longer an option. It has become a requirement. T
Dr. Richard Hackathorn is president and founder of Bolder Technology Inc., a consultancy in Boulder, CO.
© Teradata Magazine-June 2006
back to top
|
Quick Links
Print
Send to colleague
PDF
E-mail us
