Jim answers questions from fellow Drillers
(More questions with answers here, Work Overview here, Index of concepts here)
Topic Overview
Hi again folks, Jim Novo here.
Today, we’re asked for a simple definition of retention. Problem is, the data / biz model really creates the definition. Meaning, you gotta match the creation of metrics with the actual actions. So I call for segmentation first so we can put some “actionable” stuff in the mix.
Make sense? Let’s do the “simple” (easy? maybe not) Drillin’ …
Q: For an online retailer, what is the best way to gauge retention in its most basic and simplest form? % of orders that are from repeat buyers? % of orders in month 2 who are repeaters that first bought in month 1?
A: I would take direction on this from the actual results of campaigns. Basically, at the point a customer no longer responds, they have defected. Perhaps this averages 3 months or 6 months after 1st purchase, and there will be category or price segments within these “time” segments. Retention is really measured by the defection.
Now, that’s not to say that % orders from repeats or the other one you mentioned are not valid, but I suggest you think about the specific question you want answered by the metric you choose. % orders from repeats, for example, is a common metric in mail order but is often biased by campaigns, e.g. if you ratchet up customer acquisition during a single month you poison your own metrics.
You can “reverse engineer” into it by looking for people that have already defected and then find their start date, then take the average length of time between start and finish. Let’s say you agree that a person who has not bought in over 12 months is probably no longer a customer. Find all those people, find their first and last purchase dates (exclude 1x buyers), calculate average months between first and last purchase.
If this number is 8 months, then your average LifeCycle is 8 months long, and your “active” customer base is therefore everyone with a purchase in the last 8 months. Divide this number by total # of customers, and you have your retention rate.
But if you are going to act on this information, averages are not very useful. So I would further segment this group by original campaign source, category of first purchase, price point, and so forth so you begin to see patterns that can be acted on.
For example, let’s say you find new customers from PPC campaigns have a 10 month LifeCycle and new customers from Display campaigns have a 6 month LifeCycle. This is an incredibly important and highly actionable piece of information on several fronts, from allocating campaign spending to the triggering and content of customer retention campaigns.
Oh yea, I forgot, you said “simple”. OK, for simple, the “Wall Street” standard used throughout the direct retailing industry is “12 month active”. What percent of customers have made at least 1 purchase in the past 12 months? That is retention rate.
The problem with this approach is it begs for a “last ditch” retention marketing effort at 12 months since last purchase. But if the LifeCycle is really 6 or 10 months long, you will be late and the program will lose money. This is often why so many people say they can’t retain customers – they are using the wrong metrics, and acting too late to really save the customer. “Winback” is not retention.
That’s why I always prefer to match the metrics to the actions. It simply doesn’t make sense to me to know customers have an 8 month LifeCycle and then use a retention metric called “12 month active”.
You should at least use “8 month active”, if you have the resources to figure out it is in fact 8 instead of 12. The “12 month active” is used as a default because it lines up with the annual reporting of financial statements, but is often not a true measure of customer behavior – it’s simply “convenient”.
Jim
Get the book at Booklocker.com
Find Out Specifically What is in the Book
Learn Customer Marketing Concepts and Metrics (site article list)
Download the first 9 chapters of the Drilling Down book: PDF