It’s a question that comes across our desk almost weekly, “When should we send out the email, when is the best time?” – The answer as with so much in email marketing is… “it depends”. Working out when you should send an email is really unique not only to the company sending it but even unique to the content of the email itself. If your goal is to increase your engagement rates then you are going to need to work out when is the best time to send emails.
We have been sending out emails for quite a few years now and to say ‘we’ve seen it all’ would be a bit of an understatement. With over 5 million emails sent every month by myself alone. Therefore we can say with a great degree of certainty that there is no one perfect time to send an email. The best time really does vary from industry to industry, business to business as well email to email. Unfortunately for you dear reader there is no singular time that is best for all emails to be sent. Although that would make all our lives a lot easier.
The main goal of any email is to drive traffic to a website. This email engagement can only be improved if every part of the email is carefully designed to suit the audience. This includes everything from the pre-subject line, copy, design, email length, buttons and even the send time.
Test, test, and test again
Getting email engagement to increase really does require quite a lot of different tests. This includes testing the send time, subject lines, copy, design, and other key elements of the email. Ideally each aspect of the email is tested one part at a time as to not cloud the results from any specific test.
Having so many things to test and evaluate may seem daunting at first but by systematically working through each with a number of A/B tests you should start seeing patterns of engagement. Make sure to test one aspect at a time and also try run 1-3 A/B tests per item so you are sure of the results. Employing other tools like our subject line creator tool can also assist with this process.
1. Divide your list into segments
The first step is to divide you database into smaller segments. Ideally the divisions are not arbitrary but based upon matching characteristics such as, purchase history, geographic location, age, gender or as many matching characteristics that seem relevant. Hopefully by grouping similar subscribers together they will produce less random results and make testing to those segments more accurate.
Most email marketing platforms have segmenting tools built in and if not we can assist with any data segmentation you might require.
2. Create your tests
With your newly created segments it’s time to start testing. It is important to be able to measure the success of each test so try not to test multiple things at once. Always be goal orientated with each test. For example, “Does placing high value products near the top of emails result in higher sales for these item?” Make sure you tests are also based on some real world knowledge, for example people will always spend more closer to pay day. So this might skew some results if you’re testing close to those days. Try and isolate your tests as much as possible.
It is also important to also build on the findings of your tests. So for example if your Sales email is always the most profitable email and you know people spend more on payday. You should certainly then test if your Sales email is more effective if sent closer to payday.
3. Divide each segment into control and test groups
Once you have decided what you are going to test divide each segment into two equal numbered sub-segments. Your ESP should be able to do this for you. It is important to ensure each sub-segment is large enough to produce meaningful test results. If you think the segments are too small you might want to adjust what you are testing or add more data. The final option would be to run more tests to remove and randomness from the results. There is also a useful calculator you can use to calculate a good size
4. Create two versions of the email
To make the test create the email as you normally would then create a version that will test your hypothesis. This can be anything such as reordering of content, subject line, overall design, button placement.
5. Measure the results
Ideally your ESP has a robust reporting suite or heat-map capability. This should allow you to easily see which email generates more engagement as well as allowing you to see what element of the email is generating all the clicks. To really make sure of the results you could run 1-3 additional tests, testing the same thing to remove and randomness from a one off test. For evaluating send times make sure that you’re getting the same type of engagement you would expect regardless of when you send the email. Then choose the send time that gets the most engagement.
Build on the wins
Now that you have established the best time to send or any other aspect you have been testing implement these results on the main database sends. As long as the results are replicated in the main sends you are good to begin at the beginning of your testing cycle again. Testing should be a consistent practice that you continuously include into you marketing calendar. Remember also that just because a Sale email might perform well close to pay day doesn’t mean you should send your welcome emails out then to. You might find Welcome emails perform better if sent only 30 minutes after sign-up.
The key point with trying to improve engagement through email testing is to remember to constantly tailor your tests and ultimately your approach to your audience. Use your educated guesses to guide your questions and then make decisions based on the real data you get back from tests.