Tel: 01865 242425

Tel: 1-855-835-0471

Tel: 1300 993 579

A/B testing your email design – 10 tips not to overlook

Email testing

Ever wondered know how many emails your customers send and receive every day? 121. It’s no wonder people have just seconds to scan emails and decide whether they’re worth their precious time.

If your email marketing click through rates are less than rosy, maybe it’s down to how your template looks, so tweaking your email design could mean far better results.

Before we reveal our 10 tips on how to A/B test email design, a word of warning. Only test one element from this list at a time, and keep records of every result. If in doubt, read our article on how to A/B test your emails.

1. Color up your email marketing buttons

People associate different colors with feelings and behaviors. Here’s a breakdown of known color associations. Find a call-to-action (CTA) button color that works best. For example, does red convert more people when the CTA incorporates urgency or a deadline?

2. Picture placement

According to The Nielsen Norman Group, people read in an ‘F’ pattern. Keep this in mind, and play with your image placements in relation to your text. For B2C, you might want a big image with a simple CTA. For B2B you might need more wording to convey professionalism and credibility.

3. Backgrounds matter

Try out a few color and pattern options for your backgrounds. Does a seaside theme in February encourage more people to book their summer cruise? Vary your backgrounds depending on the theme and track the impact on conversions.

4. The long and short of it

If your email design looks good with lots of content and is easy to read, then run with it. It could be though, that your audience has a shorter attention span, and prefers concise, teaser emails. Compare the two styles and see which attracts most clicks.

5. To GIF or not to GIF?

GIFs are like marmite. Some people love ‘em – think they’re friendly and funny – and some people can’t stand them, finding them distracting and unprofessional. GIFs tend to perform better in B2C. Why not monitor the response?

6. Real or stock?

People might not mind a stock image, but what about creating an illustration, or even shooting a real image of your employees using your product? Test to see which appeals more to your audience. Choosing a non-stock image can improve your credibility and approachability.

7. Story styles

If you’ve got a lot of ground to cover you could represent your topics in a linear ‘masonry’ grid (like Pinterest) or as a list. If you have a longer piece, like a newsletter, it’s worth testing different styles to see which attracts most clicks.

8. Content linking

Longer pieces require direction, so sometimes, a good old table of contents can link blocks of content (with anchoring) within your email. Try a contents table and see if your clicks per article increase.

9. Spaced out

White space draws your reader’s eye to the areas that matter. But too much spacing can mean that important stuff goes below the dreaded fold. Testing different template styles, spaced and un-spaced might help to increase click through rates. Remember, always test against your standard template, rather than testing changes against each other, to get a true benchmark.

10. Be responsive

More people are opening their emails on mobile devices. If you don’t have a responsive template, you’re missing out. Dobbies tested their emails on a responsive template and saw a 400% increase in clicks. Look at your email report to see how many readers are using mobile. If it’s a low number, it might be better to focus on other areas first.

Have we sparked your interest?

Now, download our free e-guide Optimize your email campaign performance with testing to learn more about A/B testing, what to test and how to interpret the results.


Tags: , , , , ,

  • Nice list, of course there is always more to be tested 🙂 I wouldn’t think that an anchored index list to link withing the email is a good best practice though, If you have an index, best to link them to the landingpage.

    • adestramarketing

      Thanks Jordie! Nothing beats your 150+ testing ideas list 🙂 We’ve seen some companies use the anchoring successfully, but I suppose the only way to know what is best is to test.

      • From experience there are a couple of problems with anchoring inside email.

        If a user clicks on an anchored link inside the mail, this is not registered (no link tracking) unless it links to an online version of the mail. In which case, why not just send them to a landing page?

        Similarly a click on a link anchor that already is in the reading window will often not result into scrolling in an email client, making it look as if the link is broken / not functioning. So that makes for a poor User Experience.

        So if i’m making a test plan, probably anchoring would be somewhere lower on the list of testing prio’s.