Last month, I found my soapbox and climbed on up in order to dispel a myth found in a vendor-authored white paper. The myth: you can perform world class analytics without the bother of a data warehouse. From the lofty but sturdy perch on the box, I refuted two vendor-initiated misconceptions about the data warehouse. First was the contention that the warehouse is nothing more than a central copy of operational data for reporting - where, oh where is that cleansed, integrated, clearly defined, metadata-rich, subject-oriented enterprise view of data that we business intelligence (BI) practitioners know and love? Second was the claim of vastly faster delivery of analysis capabilities if you skip the time-consuming warehouse construction activities - sorry, but cleansing, integrating, defining (across business lines) and documenting data from our creaky old legacy applications takes time, and if you skip these steps, you get what you pay for: very little.


Now it is time to climb back down (too bad, I liked it up there) and play devil’s advocate. Why switch sides this month? Because, in support of their conclusions (which I still believe are erroneous), the vendor did highlight a few serious data warehouse danger signs. If undetected and uncorrected, these danger signs can cause a BI project to lose momentum, fall short of business expectations or fail altogether. The good news is that dangerous does not mean unavoidable. In fact, successful BI teams are not only aware of these issues but have turned the detection and correction of them into best practices for both BI program management and data warehouse implementation methodology.


The first hazard is failing to realize that to generate value, a BI project must be demand driven. I could not agree more; in fact, we should shout this one from the rooftops (forget the soapbox - bring on the ladder). While common sense might dictate otherwise, this is actually a fairly widespread problem. Too many data warehouse projects are initiated on the “if we build it, they will come” premise. And all too often, no one comes. How do we ensure that our BI projects are demand driven and will meet actual business need? Simple - look to business strategy and engage the business. Choosing BI projects that are tied closely with organizational strategy and facilitate organizational objectives and goals is a fine way to ensure that the data warehouse will deliver measurable value. Also, business involvement in the actual BI project is imperative, and engaging the business to help prioritize potential projects is even better. With a little guidance from the BI team, the business can rank its information needs by impact on strategy and satisfaction with current capabilities. This type of business involvement provides the BI team with a laundry list of projects that tie to strategy and provide analysis capabilities not currently available within the organization. If implemented correctly, these are sure to provide value and can correctly be termed demand driven.


Hazard two is not ensuring that the analytics you are delivering can and will be embedded into the business process. To justify this claim, the vendor touted a survey where just over 55 percent of respondents indicated that the data warehouse did not provide enough utility to motivate them to embed BI into their business process. Unfortunately, I have to concur with this one as well. Far too many data warehouse projects solicit requirements from the business, conduct the project in a vacuum and throw the results back over the fence to the business when they are done. The right thing to do is to include culture change and business process examination in the BI project methodology. Employees must be enabled and trained to use BI results to actively improve the day-to-day processes. Business units must willingly cooperate and share information. Leaders must understand what it means to benefit from analytics and set company policy accordingly. Business processes must be examined and, if necessary, realigned to handle the new flow of information.


Demand-driven BI projects based on a foundational data warehouse that result in business processes with embedded BI not only exist, but they also produce tremendous value. Consider an acknowledged BI and customer relationship management (CRM) leader, Continental Airlines. They leveraged a world class data warehouse and front-line customer service processes embedded with BI capability to transition from “worst to first” and then from “first to favorite” in a tumultuous industry. They achieved more than 1,000 percent ROI on a $5 million data warehouse investment in the first five years of operation, and they were awarded best customer service, best international service, best airline and best technology.1 In times of trouble, e.g., flight cancellations and delays, Continental gate agents know a passenger’s profitability (much different than frequent flier status), past flight experiences and upcoming flights. Operations personnel know airplane passenger loads as well as final destinations for connecting passengers and can predict downstream impacts of routing, staffing and aircraft changes. Security proactively detects fraud through booking analysis and can hone in on suspicious activity.


The combination of active business involvement, process realignment, and a well-constructed data warehouse will enable world class analytics.




  1. Continental Airlines Enterprise Case Study, Teradata, 2004.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access