tag:blogger.com,1999:blog-32202893.post8433816063749797720..comments2023-10-18T08:32:17.510-07:00Comments on Kevin Hillstrom: MineThatData: Williams Sonoma: Incremental Online Sales and Matchback AnalysisUnknownnoreply@blogger.comBlogger2125tag:blogger.com,1999:blog-32202893.post-54505712830926031462006-11-18T03:08:00.000-08:002006-11-18T03:08:00.000-08:00Thanks, Kevin, for throwing additional light on th...Thanks, Kevin, for throwing additional light on this topic.<br /><br />It is my experience that the percentage of online orders utilizing a catalog has sloped downward from a high of 95%+ some years ago to about 82% in 2004, around 78% in 2005 and, likely, 72-70% in 2006. <br /><br />It seems clear, and irrefutable, that there is a growing shift from the paper catalog to search, and that this logical shift will continue.<br /><br />One simple method for measurement which I first suggested in the keynote speech for the 1998 MeritDirect Co-Op conference was to add one required field to the online order/shopping cart (making sure the the question has to be answered before the order completes). That question is:<br /><br />Did you use or refer to our paper catalog at any point for this order?<br /><br />It may not be 100% accurate, but it will give you a reasonable read on this question fairly quickly. Regardless, it is knowledge that all catalog/online marketers--B2B, B2C or hybrid--must have in order to make informed decisions relative to channel mix budget allocations.<br /><br />Best regards<br /><br />Don Libey<br />www.libey.comAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-32202893.post-23841195026295250632006-11-17T21:54:00.000-08:002006-11-17T21:54:00.000-08:00Interesting post Kevin.
I'd personally be inclined...Interesting post Kevin.<br />I'd personally be inclined to carry out the analysis by running the A/B test as you mention; though probably taking a much smaller control group of customers who don't receive a catalogue to mitigate potential loss of revenue in case the catalogue really is that effective.<br />In the analysis of attributing all catalogue recipients vs online orders I understand the matchback approach is more commonly used - albeit (IMHO) not a valid approach for the reasons you outline. <br />If emails are also sent out, then simple tagging and tracking via the web analytics tool can correctly attribute the last causing effect for the sale. Through we then want to start considering the complexity of the events that led to the sale - a common problem with web analytics (easy to attribute a search sale or a banner sale but how do different channels contribute to the sale - the real multi channel issue!). It's for this reason that the banner 'view-through' metrics are becoming less valuable in a multi channel world. <br />Not easy to answer without understanding who are catalogue recipients..<br />Best, James.Anonymousnoreply@blogger.com