Before I get to my case for a new email marketing metric, I want to first make reference to something that was written not all that long ago. Entitled Understanding The Email 'Frequency Math Effect' it appeared on the Email Insider section of MediaPost and was penned by Loren McDonald.
I think Loren’s post was very well-written and certainly thought provoking, which is what a good post ought to do: get you thinking. And that’s precisely the effect his post had on me.
As we consider the impacts of the “Frequency Math Effect” we would be well served to remember that statistics, if improperly used, can lead us down false paths. 19th Century British Prime Minister Benjamin Disraeli may have put it best when he wrote: "There are three kinds of lies: lies, damned lies, and statistics!"
To start, remember that any rate is calculated as a numerator divided by a denominator. For instance, Click-Through Rate (CTR) is Clicks (the numerator) divided by Pushed (the denominator). As you increase mailing volume, for any rate to remain unchanged then any increase in the denominator must be matched by a proportional increase in the numerator. And in reality as Loren pointed out in his initial post the Frequency Math Effect suggests that this is unlikely to occur: ever-increasing mailing volumes will generally tend to increase the numerator, but not as quickly proportionately as the denominator.
For email marketers a reduction in a rate is typically viewed as a bad thing. There are a couple of exceptions, however: Unsubscribe Rate and Spam Complaint Rate. A drop in either of these two rates is generally viewed as a good thing. This can lead the marketer down a dangerous path, however, particularly if cadence is being increased. Follow a reductio ad absurdum example with me to see the impact:
To start, let’s call the sum of Unsubscribes and Spam Complaints “Attrits” (yes, it’s a real word that is the base of the word attrition meaning to “wear away”). Using round numbers for illustration assume you have one million customers that you mail once a week and that this mailing generates a 0.1% combined unsubscribe and spam complaint rate (1,000 attrits). I think most of us would agree that if you mailed this customer list once a day you would likely get more than 1,000 attrits (for conversation let's say 2,000 attrits are generated by this higher cadence).
But remember, this is an illustration in absurdity so what happens if you mailed the entire list once an hour? Or once a minute?! That would be 10,080 individual mailing drops (!) and I think we can be pretty safe in assuming that every customer would have unsubscribed or hit the “This is spam” button by the end of that week! Effectively, after one week we would have turned one million opt-in subscribers into one million attrits.
In the example above total Unsubscribes plus Spam Complaints would skyrocket while at the same time their matching rates would plummet (to perhaps as low as 0.01% combined). This is what Loren calls the "Frequency math effect." But there is something more going on here that is important to recognize. While the so-called "Frequency math effect" can impact any rate that you measure in email marketing, for most of those rates the numerator can theoretically increase indefinitely at the same rate as the denominator. But that is not true for Unsubscribe Rate nor for Spam Complaint Rate.
Why? Because presuming they don’t later opt back in, a customer can only attrit from a mailing list once. In my absurd example above up to ten billion emails were pushed. However, even if that had been 100 trillion emails, the numerator could never have risen above the one million records on the mailing file. As a result, as mailing cadence increases both Unsubscribe Rate and Spam Complaint Rate become progressively less useful metrics for us to rely on to diagnose our mailing program.
The Proposed New Metric
I would like to suggest that we need a new metric, one I call the "List Attrition Rate." Similar to how email marketers typically calculate both a "Click-Through Rate" and a "Unique Clicker Rate," the List Attrition Rate is designed to be a measure of unique customer behavior. However, unlike other unique metrics commonly used in our industry this one is not based solely on one campaign, but rather is calculated using all touches against the customer in a given week.
The List Attrition Rate is calculated as ("Weekly Unsubscribes" + “Weekly Spam Complaints”) divided by "Weekly Unique Recipients Pushed." The higher your List Attrition Rate, the worse the health of your mailing program. As a result this metric is intended as the “Canary in the Coal Mine” that can warn you of dangerous mailing practices. Not coincidentally this metric is explicitly intended to act as a restraining force against our natural impulses to over-mail our customers.
Consider the List Attrition Rate metric in the context of our absurd example above:
Mail once per week:
List Attrition Rate =
1,000 Weekly Attrits / 1,000,000 Weekly Unique Recipients Pushed =
Mail once a day (7 times in a week):
List Attrition Rate =
2,000 Weekly Attrits / 1,000,000 Weekly Unique Recipients Pushed =
Mail once a minute (10,080 times in a week):
List Attrition Rate =
1,000,000 Weekly Attrits / 1,000,000 Weekly Unique Recipients Pushed =
The List Attrition Rate is an intriguing metric because it measures the rate something bad occurred (the Unsubscribe or Spam Complaint) as a ratio to the number of times it could have occurred, remembering that each customer can only exit the mailing list once. Set your threshold List Attrition Rate at a low enough level and you may finally have a tool to push back against other voices in your business trying to get you to over mail your customers.
The bottom line in all of this is we marketers need to apply the one metric that is applicable to everything we do: The common sense metric. If we look at something, anything and it doesn’t make sense or seems of out whack then we probably shouldn’t do it i.e. sending over 10,000 mailings in a very short period of time as per my earlier example.
Modern Marketers must orchestrate and deliver marketing messages that are relevant to individual preferences and behavior. Getting email delivered to the inbox is critical to this process which is why you need to download Email Deliverability: Guide for Modern Marketers.