Correspondence: Tom Breur, XLNT Consulting — 'turning data into dollars', Langestraat 8-03, Tilburg 5038 SE, The Netherlands. Tel: +31 6 463 468 75; E-mail: tombreur@xlntconsulting.com

¹runs consulting firm XLNT Consulting (www.xlntconsulting.com) dedicated to helping companies make more money with their data. His fields of interest span data mining, analytics, data quality, IT governance, data warehousing and business models.

Received 2 March 2007; Revised 2 March 2007.

Abstract

Measuring campaign effectiveness is very important. After all, you need to measure to manage. Data mining has been introduced into mainstream marketing, and at the moment it is used most frequently to improve targeting. 'Closing the loop' is key in state-of-the-art database marketing. It means testing measuring–tweaking campaigns. Passes through these cycles are run at increasingly higher speeds. By manipulating both marketing execution and targeting, one attempts to increase response. Since these effects operate simultaneously, the influence they exert get mingled. As a consequence, measuring the effectiveness of campaigns is slightly more complicated. The author describes a comprehensive test-design to evaluate the relative contribution of marketing execution and data mining models in increasing response. As data mining models get reused, their effectiveness over time needs to be tracked. This framework includes both one-off evaluation and longitudinal monitoring of data mining models and marketing execution.