In my previous entry, I shared one of my favorite case studies from my book. The online retailer eBags experimented with sending messages at a unique time targeted to each recipient. The results were amazing: the company was able to increase some of its metrics by almost 200 percent.

The technical complexity of timing messages individually required the eBags team to pick just one method for determining each recipient's timing. Team members ended up choosing opt-in date. So, for example, if a customer opted-in at 2 p.m. on a Tuesday, they used the same time of day and day of week to send their campaigns. They reasoned that if someone had a specific time available to opt-in, they would have that same time available later to respond. Their results certainly confirmed this method.

eBags only got to try one method, but it hasn't stopped a lot of discussion in the years since. For example, would the timing of the last click-through or last open be more appropriate? And, how do you get around the problem of a static time? It's almost inevitable that someone's best timing in the summer will change in the fall. Schedules are going to shift. Obviously, the ideal approach would be constantly tune the correct time of day based on recent response data.

It's exciting to think about the response rate improvements marketers could drive if they had access to the open-minded technical team that eBags did. It's even more exciting to think about all of this operating at the press of a button -- no more work than scheduling a traditional mailing, except that the email system takes care of each recipient's individual time and constantly fine tunes it.

It would be nice...