AARP, Inc., formerly the American Association of Retired Persons, is a U.S.-based membership and interest group, founded in 1958 by Ethel Percy Andrus, PhD, a retired educator from California, and Leonard Davis, founder of Colonial Penn Group of insurance companies. As described in their Web site (aarp.org), AARP is a nonprofit, nonpartisan, social welfare organization with a membership of nearly 38 million that helps people turn their goals and dreams into real possibilities, strengthens communities, and fights for the issues that matter most to families—such as healthcare, employment and income security, and protection from financial abuse.
A Growing Demand for BI
In 2002, the organization first launched a BI initiative that would centralize information (AARP has offices in all 50 states as well as the District of Columbia) and empower its staff with current, relevant, accurate, and flexible analytics to:
Match services and product offerings to membership base and expectations.
Improve member profitability, retention, and acquisition.
Protect AARP brand image by managing relationships with third-party service providers.
This insight helped fuel AARP’s success and, with this success, came larger data volumes and an increased demand for new analytics.
By 2009, the BI team faced a new challenge. Its data warehouse—based on an SQL relational database from Oracle—could no longer keep up with the demand. The team experienced more than 30 systems failures that year. This was both unacceptable and costly.
System performance was a key concern as well. As the data volumes grew, daily loads into the warehouse couldn’t be completed until 3:00 p.m.—which affected how long staff had to wait for reports. “Our analysts would run a report, then go for coffee or for lunch, and, maybe if they were lucky, by 5:00 p.m. they would get the response,” says Bruni, Practice Director, Business Intelligence, AARP. “It was unacceptable. The system was so busy writing the new daily data that it didn’t give any importance to the read operations performed
Analysts also couldn’t create ad hoc queries without IT intervention. When IT received a request for a new type of report, the BI team would have to optimize the queries and send a report sample back to the requestors for review. The process, from start to finish, could take weeks to months. Finally, with more than 36 terabytes of data in the data warehouse, staff found it impossible to back up the system each night. Backups were limited to a few critical tables, making it difficult for staff to create an effective disaster recovery plan.
According to Bruni, if left unsolved, these challenges could have affected AARP’s work. “Analytics provide key metrics that are critical to evaluate how well our membership and social goals are being attained,” says Bruni. “It is essential to enabling continuous improvement and decision making to support member needs.”
Creating an Agile BI Environment
As Bruni’s team looked to modernize the BI environment, they evaluated two options—upgrading the existing environment or moving to a single data warehouse appliance. “We found the cost of each option comparable, but only the appliance provided us a paradigm shift in terms of the performance we needed,” says Bruni. “Among the different partners we looked at, the IBM Netezza data warehouse appliance provided the safest bet because it didn’t require the data model fine-tuning that other data warehouses do. We were also able to try the solution before we bought it to see whether it really could do everything we needed. Most vendors do not provide this type of ‘try-before-you-buy’ option.”
In building the new environment, the organization adopted a “Scrum” development model, usually used by software developers, to provide a framework that shortens development cycles and speeds time to market for BI requests. “Using Scrum in data warehousing is kind of unheard of,” says Bruni. “But the basic premise it provides is an agile, iterative process that enables us to rapidly transform our users’ analytic needs into operating reports that show meaningful data.”
Within 9 months from the acquisition of its new platform, the team had converted all the scripts and procedures from Oracle Database into the IBM® Netezza® data warehouse appliance. Core accounts and membership data (which resides on an IBM DB2® for z/OS® database running on an IBM System z® server), financial and human resource data from other smaller databases, and campaign analysis and segmentation data from third-party data sources are now loaded in the IBM Netezza data warehouse appliance nightly and accessible via the organization’s BI tools without interruption.
Running Complex Queries at Lightning Speed
In terms of performance (which was the BI team’s most pressing concern), daily data loads are now completed before 8:00 a.m.—a 1,400% improvement—and reports that previously took minutes to run are completed in seconds—a 1,700% improvement. The solution also helped compress the data size from 36 terabytes to just 1.5 terabytes, enabling staff to easily back up the data warehouse in only 30 minutes.
Equally important, the nearly 220 human resources, finance, marketing, and campaign staff members that use the system can now conduct what Bruni refers to as “train-of-thought analysis”— creating ad hoc reports to test theories regarding membership needs. “The IBM Netezza data warehouse appliance is like driving a Ferrari,” says Bruni. “We have opened a whole new realm of possibilities to our internal customers, who are actually able to create reports on-the-fly and get the results back in a matter of seconds. In the first few months of operation, we saw a huge spike in the number of reports being created—nearly three times the number that we had previously supported. With the deep dive they can conduct now, we’ve seen a steady growth in member renewals, acquisitions and engagement.”
Achieving Rapid ROI
The new platform has also enabled the organization to redeploy IT support staff from the BI group to other areas. Previously, the team needed one full-time database administrator (DBA) along with part-time support from the organization’s storage area network (SAN) and midrange service teams. “It’s amazing,” says Bruni. “We no longer need IT support. The IBM Netezza data warehouse appliance is shipped already optimized. Give it power, give it network, and you’re done. It doesn’t need anything else.”
These improvements have enabled the organization to realize a 9% return on investment in the first year, with an anticipated 274% ROI by the second year, and a 347% investment by the third year. “Our initial analysis projected a positive ROI already in the first year—which is very unusual for infrastructure upgrades given all costs are incurrent in the first year,” says Bruni. “Our actual ROI post-implementation was even higher as we completed the swap three months ahead of schedule.”
Expanding the Influence of BI
By modernizing its infrastructure, Bruni’s team has elevated the value and perception of BI in the organization. “After we moved to IBM Netezza, the word spread that we were doing things right and that leveraging us as an internal service was really smart,” says Bruni. “We’ve gained new mission-critical areas, such as the social-impact area which supports our Drive to End Hunger and Create the Good campaigns, based on the fact that we have such a robust infrastructure and that we changed our approach to business. We can develop in a more agile way from a development standpoint. From a program management standpoint, it shrinks our release cycles from months, which is typical with traditional data warehouse infrastructures, to just weeks.”
Questions for Discussion
What were the challenges AARP was facing?
What was the approach for a potential solution?
What were the results obtained in the short term, and what were the future plans?