Publications

Blue Martini recently completed a large case study using business intelligence and web analytics at Debenhams, Britain's Favourite Department Store.
The article will interest the KDD community, especially those doing web mining.
The case study highlights many interesting findings from their web site, that are deeper than what is typically done in web analytics:

• Visits that included a search generated 1.5 times as much revenue as visits that did not
• E-mail campaigns generated 4.8 pence per e-mail and 52 pence per clickthrough
• Shopping cart abandonment rate of 62% prompted a revision of session timeout duration
• Product associations can help merchandisers make better cross-sell recommendations and react faster to new trends.
• Recency, Frequency, Monetary (RFM) analysis revealed segments that should be targeted
• 53% of orders were driven from the wedding list (wedding registry)
• The catalogue link, which allows visitors to type an item number from a print catalog, generated more revenue than the women's link, which attracted seven times more visits
• Search analysis revealed interest in products not available online, leading Debenhams to direct visitors to the nearest store where those products are available
• Referrer report shows significantly high revenues associated with visits referred to by AOL
• Product associations can help merchandisers make better cross-sell recommendations and react faster to new trends.
• Bot/crawler visits accounted for 40% of sessions to the site, skewing many metrics unless removed.

Ronny Kohavi VP, Business Intelligence Blue Martini Software