REVIEWER: Jerry Thompson, senior vice president and chief information officer for Credit Union of Texas.

BACKGROUND: Credit Union of Texas is one of the largest credit unions in the country with more than $1.1 billion in assets. It operates 11 branches serving more than 155,000 members in the Dallas-Fort Worth metroplex.

PLATFORMS: We run Unica Corporation's Affinium Model on DB2 Universal Database Enterprise Edition V5.2, 7.1 and 7.2; DB2 Warehouse Manager V5.2, 7.1 and 7.2; one Netfinity 7000-M10 with 2GB memory, four CPUs, two ServeRaid Adapters and 500GB of IBM disk storage; and four Netvista Workstations with 1.8 MHz Pentium 4 processors and 2GB of memory.

PROBLEM SOLVED: We selected Affinium Model to improve our understanding of members' current and future needs and to enable highly targeted, personalized marketing campaigns.

PRODUCT FUNCTIONALITY: Since implementing Affinium Model in October 2000, we've found it much easier to use than any other package. As I like to say, "If you can use Excel, you can use Model." We also will be finding additional creative ways to apply modeling to our marketing initiatives in the future.

STRENGTHS: Ease of use is the main strength of the product. Our marketers have used Model without requiring statisticians or IT support to create 15 successful models. For example, for a leasing campaign, marketers used Model to determine that owning a truck or the presence of a child over the age of six were among the strongest factors predicting whether or not an individual would lease a vehicle through us. Using Affinium Model, we targeted only 15,000 prospects but achieved a 4.5 percent response, which yielded $510,000 in annual revenue. Automation is another strength; specifically, Model's ability to automatically find the best algorithms for a particular situation without trial-and-error searching. For example, we used Affinium Model to predict new branch locations and achieved profitability within 90 days rather than the usual one year. Using cross-validator and propagation algorithms to analyze our checking account customer data, Model predicted that branches that were within a 10-minute drive would attract new accounts and those beyond would not. Affinium Model also ranked the locations based on projected profitability.

WEAKNESSES: SAS modelers need to be trained on how to use the modeling algorithms in Affinium Model and may require extra time to transition their analysis to Model.

SELECTION CRITERIA: We needed a solution that was sophisticated yet fast, efficient and easy to use. We found Affinium both robust in functionality and extremely user friendly. Affinium Model has helped us automatically identify variables that we never thought would influence our models. We have plans to increase our use of Model throughout our operations because we have found its usefulness is only limited by our imagination.

DELIVERABLES: The customer valuator model predicts monetary value and profitability of individual customers over time, helping target profitable customers and allocate marketing resources cost-effectively. The market segmenter/profiler model identifies and analyzes attributes that drive customer behavior. The response modeler identifies individuals most likely to exhibit certain behaviors based on purchases, demographics and lifestyle data. The cross- seller model prioritizes purchases for each customer and forecasts what he or she is most likely to buy next. Affinium Model also generates more than 15 reports including modeling summary, product clusters, model sensitivity summary, model variable sensitivity, model performance and model details.

VENDOR SUPPORT: Unica consultants and trainers provided thorough training and implementation support. Their help in creating a model to predict branch locations allowed us to quickly get up and running with a live application. Their willingness to give us guidance in finding creative ways to use Model to solve real business issues has been the best part of working with Unica. They are truly committed to our success.

DOCUMENTATION: The documentation is comprehensive and understandable. After a few days of practice, relying on the documentation enabled us to create our own models and test which were most effective.

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