Of course, it’s not that
simple. Every email marketer knows that sending more messages per subscriber can drive higher unsubscribes and spam reports — and, in some cases, can cause deliverability problems. Most often,
frequency is a historical choice made a long time ago and then forgotten. Many companies try a one-size fits all suppression policy based on recency of engagement.
Although many companies have thought about optimizing frequency, most struggle with determining how to even start this goal.
After looking at this problem for over a year with a large number of clients, we’ve determined the following solution:
1. Send more mail to some
2. At the same time, send less mail to others
3. Send different kinds of mail
Send more mail: Some subscribers really, really love email. If you send them more mail, the total number of emails they read and purchases they make increase. Although their read rate declines and their unsubscribes and complaints increase, the increased revenue from sending this group more email overwhelms the increase in list churn. Send this group more email.
Send less mail: Some subscribers are less enamored with your email. Although you may drive a few more opens, clicks, and conversions with more email over the short run, the longer-term increase in list churn more than overwhelms the short-term benefit. So send this group less email. These subscribers may prefer other channels, so send more messages through mobile and social.
Send different kinds of mail: Not all email program types are created equal. Triggered and contextual messages frequently drive over 10x the revenue per message that typical campaign- based mail provides. The impact on list churn is minimal from triggered programs. Many online retailers use mainstay triggered messaging programs such as cart abandonment and abandoned browse. However, outside of online retail, use of triggered messaging is less common. Even online retailers use a relatively narrow palette of triggered messaging campaigns. Back-in-stock, price change, new inventory, and other programs are less common. Increased volume from triggered messaging is “free” revenue.
The benefits from this approach are huge: In the experiments we’ve been running with clients, we have seen revenue dramatically increase by optimizing frequency. Often, we end up sending a little less mail in the near term (a few months), generating less revenue. However, the compound power of having a larger list and being able to send more mail to targeted segments can rapidly improve mail performance. There are added benefits from adjusting cadence. Deliverability problems appear to be less common when frequency is appropriately adjusted. By focusing on retaining subscribers during non-peak periods, you can benefit tremendously from going into peak periods with a much larger list of customers when they are ready to convert.
Engagement buckets don’t get the job done: How do you determine which subscribers can take more mail? It’s not as simple as looking at “engagement” buckets (subscribers that haven’t opened or clicked in 30/60/90 days). Engagement buckets don’t do an optimal job of interpreting termination rates, nor are they particularly predictive of how opens and clicks will change based on increased or decreased volume.
Since most marketers haven’t varied cadence much historically, we have found that bringing in data from a panel of millions of mailboxes allows us to build models that correctly predict what will happen when marketers change cadence with specific kinds of subscribers.
For most online retailers, now is the time to start: Since a lot of the value from frequency optimization comes from lowering termination rates (complaints and unsubscribes) over time, the sooner you get going on optimizing frequency the better. Since many online retailers are focused on November and December sales, starting this effort in May/June is important.