As we have all learned over time, a recent customer purchase from your catalog is the strongest indicator that the customer will respond to your next mailing. In fact, purchase recency and frequency dwarf all other factors as predictors of future response; that is, until they don’t.
As RFM data ages, it becomes less likely to accurately predict future purchase behavior. Specifically, as we look at the bottom segments of our circulation plan, RFM data alone is unlikely to optimize the effectiveness of a mailing. We can screen those older names through Co-op models, using RFM data from marketers that “look like us” to differentiate some of our borderline segments. This continues to be an effective approach, and CohereOne uses Co-op optimization models with most of our circulation clients.
But there is more that we can do to improve the quality of our reactivation mailings. Thousands of shoppers browse our websites every day, many of whom have not purchased from us for quite some time. Why not use that browsing data to enhance the quality of our reactivation efforts?
So how do we do this?
- Establish a browsing behavior database, with a record for each unique visitor to the website,
- In this database, collect as much data as we possibly can about each visit,
- Match the visitors to our customer list, so we can identify customer number and postal address for as many visitors as possible,
- Each time a mailing event is imminent, examine the browsing behavior and identify the shoppers that are truly engaged on the website,
- Match the engaged browsers against the list of house file names that we are not planning to mail, and send a catalog to all the matches,
This really works, and why shouldn’t it? Shoppers who aggressively browse our website are demonstrating intent to buy, and a well-timed mail piece will help our chances of securing that purchase.
We recently helped a client test this approach in their summer and fall mailings. For both mailing events, reactivation segments were selected using a transaction based reactivation model. CohereOne then used our NaviStone„¢ product to identify additional reactivation candidates based on recent web browsing activity on the client’s website. As you can see in the table below, our client was able to increase their reactivation circulation for the two seasons by about 40%. And, the additional “intent” based names performed 35-40% better than the reactivation segments that were originally identified using transaction history.
Since only 3-5% of website visitors make a purchase at the time of their visit, there is a wealth of important intent data left behind that can help us improve many of our marketing programs, both print and online.