I’ve been working on a list analytics project recently and that’s resulted in me thinking of, among other things, universes. I’ve started to think about model performance and how the changes to the economy might be impacting marketing results.

However, I wasn’t tempted to write this post until two things happened yesterday. First, I received MediaPost’s Email Insider newsletter on the subject of customer lifecycle. Then, I needed to drive to a client yesterday and noticed that I’ve been driving a different car more frequently.

It dawned on me that I’ve become an x minus 1.

I did some quick analysis and found that I drive about 1,500 miles per month for work and pleasure, pretty consistently every month of the year. I drive two cars. One is a more “fun” car and the other is my weekend car. Both get just about the same gas mileage. The difference is that the “fun” car needs premium gas and the weekend car takes regular. As I write this, regular gas at my local station is $2.979 per gallon and premium is $3.259.

Normally, I drive whatever car makes the most sense for me and don’t worry about the cost of gas. But something has changed recently (must be the economy and recession getting to me) and, while I’m buying about the same quantity of gas, I’ve switched the distribution of spending quite heavily to the lower-priced, regular fuel.

So if you were to look at me in terms of my lifetime value (LTV) for that oil company, I’ve moved down a model rank or two, at least for some period of time.

Hence my status as an x minus 1.

And I’m not the only customer of the oil company who’s done the same thing. The composition of their customers has changed. Due to the economy, some customers–like me–have likely moved down on LTV models, while others have moved up or stayed the same.

Marketing implications

Your models, whether they are acquisition, LTV, cancel propensity, retention or what have you, were likely created before the recession hit. If the universe of customers has now changed–even slightly–some inappropriate marketing is likely to happen. The possible impacts:

1. Trigger programs. Marketing actions triggered by customer model rank might contact customers with inappropriate offers, too frequent or not frequent enough contacts, driving ROI down.
2. Reactivation/winback efforts. Programs that select or suppress a certain amount of deciles, quintiles or vigintiles will begin selecting/suppressing from a universe that has changed, generally causing response and backend performance to go down.
3. Acquisition efforts. The most costly impact, because you might be targeting the wrong audience. And acquisition marketing ROI is particularly sensitive to model drift.

If you haven’t re-validated your models recently, you should check them and analyze the performance over the last six months or so. I suspect there’s been some subtle drift in the performance. And a word of warning–sometimes you might see performance drifting upward when you look back. That’s just as bad a sign, as any strange or unpredictable performance means something has changed in your modeled universe and will require action on your part.

Further, you should be mailing/emailing/OBTM to model nths at least a couple of times per year to take advantage of universe changes. Even in a recession there’s an opportunity to improve your marketing ROI through constant model improvement.

Summary and takeaways

1. Check current model performance. Does it still rank and predict customer performance? Has performance deteriorated slowly over time?
2. Consider launching a new model nth. If you haven’t mailed/emailed a model nth recently, you should consider doing so now. It’s tempting in slow or recessionary times to be conservative, but this is not the time to hunker down and cut back on modeling.
3. Consider changing up product and offers. Recessions are a great time to re-think all elements of the marketing mix. While you’re looking at your models, think about pricing, creating new products or units of sale. And don’t forget higher prices. You can–and I’ve done it–raise prices and profits in a recession, by looking for pockets of customers and prospects who’ve moved up in model rank.