Cohort Analysis - Chapter 2

The Snapshot Approach

Let’s say you wanted to evaluate the effectiveness of your retention marketing over time. How would you go about it?

One common approach would be to take a snapshot of your current customer base, then compare it to a snapshot at a different point in time. For many retailers, revenue per user is a critical metric of retention and engagement. Using the snapshot approach, you might ask: how does the average revenue per user (ARPU) in September 2013 compare to what it was back in September 2012? Or in other words, how much money was a customer spending on our site on average in September 2013, vs September 2012? If ARPU is higher now, the retailer might conclude that their retention program has been working well, since it resulted in increased ARPU.

But there’s a snag. We know that if we were to watch the same fixed group of customers over time, they will almost always “fade away”: only some subset of the group that initially made purchases will repeat, and over time even some of those repeat customers will leave for competitors or drop out of the category altogether. This is a natural component of the Customer Lifecycle (see Lifecycle Marketing course on Custora U).

The problem with comparing static snapshots is that the composition of your customer base is changing over the time period in question - and metrics like ARPU don’t take the “age” of your customers into account. Specifically, if the mix of new and older customers is changing, it’s bound to return skewed results.

For example, imagine that you acquire a lot of new customers in September 2013 through a combination of Labor Day and back-to-school promotions. All those new customers are bringing in a lot of revenue, so your ARPU will probably be quite high. But your existing customers may not be worth any more in September 2013 than they were in September 2012 - and in fact, they may be worth even less if you don’t have a strong retention program in place. ARPU doesn’t control for how long customers have been around - it just takes total revenue and spreads it out over the entire customer base.

Why is this a problem? New customer acquisition can mask retention problems. You may have a “leaky bucket,” where fewer and fewer of your customers are repeating - but if your rate of new customer acquisition is high enough, your ARPU will remain high.

 

Cohort

Back to Custora U Library