Rhenania Optimizes Its Mail-Order Business with Dynamic Multilevel Modeling
The Problem
Rhenania is a German direct mail order company. In 1996, Rhenania's CEO confronted multiple challenges: a sustained trajectory of declining sales, declining market share, and declining profits. The company downturn had occurred even though Rhenania had followed the standard marketing approach to managing customer contacts and choosing the best customers for each mailing. While Rhenania had increased the productivity of its separate, single mailings, its customer base was shrinking. The company turned to analytics expertise in optimization and strategic planning to increase its customer base and to increase the firm's profitability.
The Analytics Solution
Rhenania's marketing director had a background in analytics modeling. He recognized that mail-order companies' traditional practice of optimizing individual mailing productivity is actually suboptimal, since it leads to shrinking bases of active customers (customers who ordered in the last 12 months) and lower profitability over time. He convinced the company's new CEO to try an analytics approach that went against conventional wisdom.
The analytics team developed a dynamic multilevel modeling (DMLM) approach to determine the optimal frequency of catalog mailings, optimize mailings according to customer segments, and determine when customers should receive reactivation packages as opposed to catalogs.
The Value
Rhenania turned its catalog mailing practices around within one year and elevated its market position from fifth to second place in Germany. The approach was so effective that Rhenania acquired two competitors, including a subdivision of the major publisher Springer Verlag.
Wrote Frederick Palm, CEO, "Today DMLM is fully implemented and every address we are mailing to is chosen by this algorithm. Since the implementation, we are doing extremely well in contrast to most of our competitors. Now we are gaining market share out of our own business and very recently we are gaining market share because of acquisitions.... In addition to these dramatic economic improvements based on this model, it is an excellent forecasting tool. We are able to look 12 months in the future to see the development of active customers, sales, and profit."
