Modern marketers love December. It's so "easy" to acquire customers. I worked with a "normal" brand who acquired 40% of their new buyers every December.
They didn't, however, elect to measure if this was a good decision.
Remember our data from yesterday? The best newbies had a 58% chance of buying again (decile one) and the worst newbies had a 28% chance of buying again (decile ten).
This graph shows, by decile, the fraction of new buyers acquired in December. Tell me what you observe:
Look at deciles nine and ten. These are the worst deciles ... and anywhere between a third and half of the first-time buyers in these deciles are acquired in December.
In other words, if you acquire a customer in December, the customer (for this brand) is generally a low value customer, one unlikely to buy again in the future. You better collect all of your profit upfront with this customer.
In the Logistic Regression analysis performed for this brand, I learned the following:
- January / February / March / April / May / September new buyers perform at or slightly above average.
- July / August newbies perform above average.
- October newbies perform below average.
- November / December newbies perform well below average in the future.