As email marketing experts, our clients often ask us how many emails they should send per week/month/quarter to get the best results. But there is no “one-size-fits-all” answer because optimizing email send frequency depends on how your list responds to your mailings, and how relevant, timely, and expected your mailings are, as well as a host of other factors.
In fact, the MarketTraq™ motto of “relevant, timely, and anticipated” is more appropriate to optimizing send frequency than any other aspect of planning an email marketing campaign. Any attempt to optimize your email campaign should begin by examining each of these three critical factors:
In short, is your message worth reading to your customers? Before running any kind of send frequency optimization, consider if you really have enough news, content, or promotional content that’s really relevant to send that will be of interest to your subscribers. If you find yourself stretching to generate enough content, consider delaying the send rather than sending non-relevant information.
You want the right message to get to the right subscriber at the right time, otherwise all the email frequency optimization in the world won’t produce great results. A large part of getting the right message to the right person is segmentation, and if you have kept up with segmenting your list properly you should consider the possibility that different segments will respond to different send frequencies.
Your subscribers likely have an expectation of when your emails go out. Often, this comes from the language on the signup form (“Click here to get our monthly newsletter” or “Subscribe to our daily tips email”).
Other times it can be based on your behavior, with subscribers getting used to emails at regular intervals. Breaking these expectations, either with too-often emails or too-rare emails can cause a large impact. So remember to test your way into the answers, rather than assuming you have the right frequency and using that. Start modest, build frequency and measure all the impacts of doing so (conversion, total sales, click, open, unsubscribe).
So now that the basics are taken care of, you can begin answering the question of the hour: “how much email is too much email?”
If you look back, we recently wrote about an interesting study that showed open rates are not negatively impacted by increasing send frequency. It’s important to keep in mind that aggregate statistics are not accurate when applied to unique segments. The best way to find out how much is too much is to test.
1. Define the Relevant Segments.
In addition to behavior or value based segmentation common in retail, or decay/aging segmentations common in B2B lead lists, the next thing you’ll want to layer in is test segment and a control segment.
How many people should be in your test segment? It really depends on your list size. Try this sample size calculator (http://www.surveysystem.com/sscalc.htm) with your total list size as the population. If you want to be extra sure, use a bigger test segment. If you want to test multiple send frequencies simultaneously, use multiple test segments.
2. Decide what you’re testing.
Ultimately, the test needs to be calibrated in such a way that the only thing that changes between samples is frequency. This means keeping your content, offers, subject lines, and send time/day identical. This isn’t really feasible for this type of experiment, since you’re going to be sending 2 or 3 more creatives to the test segments for each one you send to your control. A good solution around this is to take your test creatives and compile them into a digest that you send less frequently to your control group. This way, you keep as many of the variables identical as possible.
A key factor to keep in mind is that when testing email send frequency, you look at broader business objectives as opposed to only open or click-through rates. Have a good metric in place for judging email ROI before you begin your experiment. Why? Well, it might be that your more frequent sends decrease open and click-through rates, as less people are willing to look at every one of your emails, but maybe the ones that DO open it end up converting more frequently (perhaps making more frequent purchases, or sharing your information more often). In this case, if you judge results simply by email metrics, you would see the experiment as a failure.
However, if you judge the results based on overall business metrics, the test might be a smashing success at increasing key business metrics such as customer lifetime value or revenue.
The bottom line: Before you begin frequency optimization testing, or any other testing, it’s critical that you’ve decided on what the success criteria is for the test upfront or the data won’t tell you much about it.
3. Run Tests Carefully!
Avoid running multiple email send frequency tests on the same segment in rapid succession.
If you want to see if your results improve by moving from once-a-month emails to once-a-week emails, and then to see if bi-weekly emails are even better, you need to set up independent experiments. One warning, if you have an expectation of instant gratification you are going to need very large randomized lists without duplication across them (that would skew and distort any test if there were recipients in both lists or test segments) and the resources to run and track them all independently. Otherwise, you should plan to go in an evolutionary pace to avoid mistakes.
Optimizing email send frequency can be a great way to increase brand engagement and other key business metrics, and can impact your email ROI in a big and positive way. As always, approach it with a well-thought-out and well-planned testing approach, and let the results guide you, instead of relying on gut feeling and publicly available averages that are published at times. Those can be useful in directionally setting up your frequency experiments, but shouldn’t be considered answers, as the aggregate tends to blind you to the behaviors and response rates in the unique segments that exist in your business and in your email database.