You probably already know this (or are executing these programs), but it is worth running a query to validate anyway.
- Segment customers by months since last purchase, through June 30.
- Add a segmentation layer ... 1 = customer visited website in June 2012 but did not purchase ... 0 = customer did not visit website in June 2012.
- Based on this level of segmentation, measure the amount of demand the average customer in the segment spent during July 2012.
- Customers who visit the website but do not purchase, especially among those with 7+ months of recency, outperform those who do not visit the website in the past 30 days by a factor of 3x to 10x.
- Kick out a hotline catalog to any lapsed, unmailed buyer who visits your website in the past month and does not purchase.
- Develop trigger-based email programs for customers who are lapsed and then visit the website, sans purchase.
- Call customers who exhibit this behavior, offering your handy/dandy promotional programs that you love to talk about.