There are a couple of situations where the Recency metric has to be “tempered” by judgment and knowledge of the business. One is where there is a strong “cycle” nature to the behavior, as with renewals of various kinds like insurance or buying machinery. Here you use the Latency metric to define the cyclic behavior and then the Recency metric to look at Engagement within a cycle.  Rather than going through a detailed explanation of this idea here, see this newsletter issue for an explanation.

The other situation is where there is a “subscription” of some kind or service billing like a utility or phone bill. In these situations, the analysis frequently centers not around revenue / spend, but around the service itself as in “Recency of Trouble call”, see this page for details.  In these cases, high Engagement is negative - customers don’t usually want to be “engaged” in service problems.

Hope that helps with my post and if you have more questions on this perhaps I’ll do another post on it, I don’t want to be accused of burying the good stuff in a comment!

As for Ron’s suggestion #1, I hear your angst about letting management define the metrics but I think Ron is looking at “any” management participation as a step forward; we can tweak the metrics later!  And to be realistic, the metrics would probably vary by the mission of the bank and any operating constraints the bank might have on executing against the metrics.

How to you factor purchase cycle or seasonal buying in your Visit Recency model (or is it relevent)? For example, if I am shopping for a new insurance policy, I will most probably come back to the site several times within my purchase decision timeframe, and go quiet for a long time before I restart the whole process.

I am not so sure I agree with Ron’s #1; management could be “generous” and too inclusive, so that there could be a bias toward positive results.

So how do you get action on this? I assume “pay for performance” is not common in the banking world, so talking about increased returns on marketing investments or IRR doesn’t get much traction? Who wants to take on more risk without a reward?

We do have this problem of the “cold war” between brand and direct marketers though, so I think it would be helpful if we could all decide on a universal platform for measuring these types of ideas. Maybe then we could go to the CFO and get some respect… that’s really the primary idea behind my post.  So maybe I could persuade you to alter your metrics a little bit (checked savings rate X times in the past 6 months, last check was X days ago) though I realize you were working with a different kind of data for this and probably could not get there…

I wonder how many bankers themselves have old accounts they just let sit there for whatever reason, but if you asked them, they wouldn’t say they have a “relationship” with that bank. Yet when they get to thinking about their own business, somehow they forget and consider people with “orphaned” accounts just like their own to be “retained”.

And bankers are probably one of the best cases…at least at the bank there is some kind of asset, some “excuse” you could call a relationship. This same situation is much worse in retail and a lot of B2B, where “customers” can have no engagement for a decade and still be considered a “customer”.

Many banks will develop predictive models to guess who will close out accounts over the next year (or whatever time period) and push out messages to “retain” the account.

My argument is that this is wasted effort. And I couldn’t agree with you more that banks need to change their definition of retention.