© DecisionWorks Consulting, Inc. All rights reserved.
Print Friendly

Most organizations implementing a new data warehouse/business intelligence (DW/BI) environment are replacing or “sunsetting” a legacy analytic/reporting system. This environment may be an older data warehouse, a single or series of departmental data marts, or a collection of analytic/reporting environments cobbled together using tools such as Access and Excel. Some may be officially sanctioned platforms; often they are shadow reporting environments crafted by a business unit. Regardless of the flavor, many business users rely on these environments for their reporting and analytic needs. Elimination of these disparate systems is a common goal of new DW/BI initiatives.

Elimination of these legacy/shadow systems often provides a significant amount of the hard dollar justification for the new DW/BI world. These savings result from the proposed elimination of hardware, software licenses, and associated maintenance and support costs. The hard dollar savings often include the manpower costs involved in reconciling data across multiple analytic platforms. In addition, there is often soft dollar justification based on the value of better integrated data, greater management confidence, and improved ease of use associated with replacing the legacy environments.

Unfortunately, sunsetting the legacy environments seldom occurs in a timely manner or often not at all, placing these justification dollars at risk. It’s not uncommon to see an older legacy reporting environment running side by side with the new DW/BI environment far longer than was ever intended. Clearly, this is undesirable; if both environments continue to run, the hard dollar justification will never be recognized. At some point, management is going to look askew at this situation and begin questioning the integrity of the team responsible for the new DW/BI environment, potentially putting future funding at risk. In addition, most users will never embrace the new environment as long as the old environment continues to be available. Why change if you don’t have to? As a result, there will continue to be data reconciliation challenges due to the multiple, inconsistent analytic and reporting platforms.

It is vital to the ongoing success of the new environment to avoid this situation. The DW/BI team needs to keep its eye on the goal, ensuring that each iteration helps move closer to sunsetting some portion of the legacy environment. At some point, it is often just a matter of having the courage to actually turn something off that is no longer needed. Before sunsetting an older environment you must completely understand the requirements that must be supported by the replacement DW/BI system:

  • Are all of the data and capabilities in place to support the current reports/analyses?
  • Are there other non-analytic functions that the older environment is supporting?
  • Does the existing system provide certain capabilities, such as auditing and compliance, that are not available elsewhere?
  • Are there downstream environments dependent on data from the legacy environment?

All of these requirements have a valid claim against the existing environment. Thus, you can’t sunset the environment until you can support these requirements. At a recent client, the old analytic platform included a series of daily replicated copies of the production source. The primary purpose of these tables was to support the ETL change data capture process required to populate the environment. However, unbeknownst to the DW/BI team, some key users had obtained access to the replicated tables. They included these tables as a critical component of several data governance, data quality and internal audit processes. When the DW/BI team attempted to shut down the old system, they found they couldn’t as they had no alternate capability for supporting these requirements. Similarly, an analytic platform may find itself feeding other downstream purposes it may be totally unaware of. Again, be sure you have alternate solutions in place to support these requirements.

Once you are confident you have addressed the legacy system’s support of analytics and reporting, non-DW/BI usage, and downstream application requirements, it’s time to turn that system off! Of course, you may want to keep it processing for a few weeks while you shut off all external access and feeds from the system just to make sure you’ve got everything covered. Clearly, you’ll want to get past a month end close to see if anything breaks.

Once you’re comfortable you’ve got all your bases covered it’s time – mix up some margaritas, pull the plug, and enjoy the sunset!!

Share this:
Share with your friends

Share with your friends

Share with your friends