By Jessica Best, Digital Marketing Evangelist for emfluence
Pop quiz: What’s the only way to ensure you’re always increasing your return on investment for email marketing?
If you’ve been around the block, figuratively speaking, you already know what I’m going to say. TEST something with every email campaign. Learn something about what works with your subscribers in today’s marketplace with your product. Sure, Saturday at 5:00AM works for a body products manufacturer, but will it work for your business-to-business service provider?
Email marketing experts are always asked what day of week work best? How about the color of the buttons/links? Is 3 “pages” too long for my newsletter? Most pros worth their salt will admit: those answers can vary for every audience or product. Older audiences still like to see the words “click here” with an underline to know where to click a link, but younger audiences may be hip to a more creative call to action.
First things first: Be sure you’re randomly segmenting your contact list for all tests. Your email marketing platform may have a random segmentation tool for you to use to help split your list up in a unique way every time you test. This helps ensure you’re getting scientific results.
While we’re talking about scientific, you’ll want to be sure your segments are about 10,000 contacts or more. Smaller segments make it too easy to have big percentage swings between versions, that aren’t statistically significant. If your list or segments are smaller, use a statistical significance calculator to be sure your results are “real.”
What are some of the things I recommend most email marketers should test? A quick list of Things You should Be Testing to get you started:
1. Time of Day/Day of Week
This is one of the biggest questions I hear when I’m at email marketing conferences: “When’s the best time to send?”
The answer is the ever-frustrating: It depends.Restaurants tend to send at “hungry times” – before lunch and before dinner. Retailers like sending early in the day to catch shoppers before they head out for the day. Business-to-business sales teams usually stick to Monday through Friday. But the truth is: every audience is different. You’ll only know what works best for your product or service if you test.
2. Subject line length
You can use up to 85 characters for most inboxes… on a desktop computer. Mobile readers are only seeing about 40 characters of the subject line in their inboxes according to StyleCampaign (Check out the infographic here.)So what works best for your audience?
Just like in other channels, your tone should be defined by your company, the target audience that you want to reach and your position in the marketplace. Yes, online content can be a little less formal than offline… if it suits your brand. Play with a few “voices” that fit your brand to see which resonates with your audience best.
How graphically heavy your email is can be tested for effectiveness. If you have an audience that largely has images turned off or needs a larger amount of explanation about your product or service, words may be a better way to spend your space.
Test whether adding a fun bit of movement to your emails – rotating the image of your product to see it from multiple sides, or adding a flickering flame to that birthday candle – will drive up engagement. See how one of our clients, Rush Wade 2, used an animated GIF to drive clicks through to their portfolio video.
Don’t misunderstand: I don’t mean test if you should use personalization, but rather, how you should use it. The consumer is only peripherally aware of how much data we marketers really have about them. When we use it to be useful, we can win with (much) higher return on investment! When we overuse it… we become, well, creepy. And that can have the opposite effect on our bottom line.
No one knows for sure what that line between useful and scary is. The truth is, it’s changing every year. As data becomes somewhat “de facto,” consumers’ comfort level with seeing their own data in marketing increases, like tailoring content based on their past shopping choices, online behaviors and even cross channel purchases.
TIP: You can offer a suggested item without specifying that you already know it’s exactly what they’re most likely to buy.
TIP #2: You can personalize to the user rather than the buyer and still see improved results, like this Marketing Sherpa case study about using “doggy data” instead of the pet owner data in email campaigns.
7. Subject line content
This is a big question in the email marketing industry in the past year: What works in subject lines? There couldn’t possibly be a magic bullet for subject lines: discounts may not be part of your brand image. The word FREE may get you stuck in a spam folder instead of the inbox if you overuse it.
More importantly: what is your value proposition? Since that changes between companies, products and even campaigns, there’s no one way to write a successful subject line. Test highlighting different articles and see which has the best draw to open rates. Test wording the offer differently. You can even try addressing the reader’s pet or kiddo’s name. Find what works best for you.[su_note]FREE DOWNLOAD: Click Here to download a summary of all 7 things you should be testing so you don’t FORGET![/su_note]
That last test can be tricky. Here’s my biggest piece of advice for testing in email marketing: Make it actionable.
Set your tests up so that you’ll learn something that is actionable for future campaigns. You may split test two different subject lines, but without having an intentional attack on the subject lines, you’ve only learned which subject line worked in that case.
That brings up a final, very important point: Don’t assume your learnings are true forever. The consumer mindset, the marketplace and even your product are destined to change. Test again every year or two to be sure your results are still relevant.
So, how are you conducting your email marketing campaign testing? Before you test, make sure you’re:
- Using a random segmentation methodology.
- Using at segments of at least 10,000 or calculating the statistical significance of your results.
- Creating a test where the results and learnings will be actionable.
- Testing regularly to reprove results!