This article will highlight ways to do A/B testing on your marketing emails with an assist from solid AI prompts. ChatGPT and other AI writers are not perfect but can save you a ton of time writing emails. Especially if you are taking advantage of A/B testing.
For example, it can take a human writer a while to come up with 20 subject lines to test, right? An artificial intelligence writer can have 20 written in seconds (depending on the topic and quality of AI prompts from the human).
So it may be more efficient to let the bot create options then let the email marketer choose the best two subject lines out of the 20 to edit then test.
Let’s look closer at A/B email testing and how AI writing can improve the testing process.
A/B testing emails is defined as testing two versions of one email to see which version performs best. The test is performed on part of your email list then you check your ESP analytics to view the results. Then, you can send the email that performed best to the remainder of your list.
The idea is to gain an edge in email engagement via testing. 20% more opens would be great. But guess what. 5% more opens makes a difference too. It just depends on the size of your list and even the cost of your products / services.
Key point: A/B testing doesn’t have to produce giant leaps in clicks or conversions to be worthwhile.
What to test and how to use AI prompts to speed up testing?
Below are 12 items you can split test in email campaigns. We’ll break down each one after the list:
- Subject line
- Preview text
- Email timing
- From Name (sender of email shows in readers’ inbox)
- Text-only versus adding images
- Personalization
- Length of email (also sentence / paragraph length)
- Calls-to-action
- Email format testing
- Cross-selling and upselling
- Transactional emails
- Human-written versus AI-written
Email Subject Line with AI Prompt Examples
If you test nothing else, the subject line is a must for A/B testing emails. It’s easy and shouldn’t consume a ton of time. An AI writer can cut down your time spent on producing subject lines. You just need good AI prompts.
I got these six example subject lines using the Rytr AI prompt “write a subject line for an email about an email conference event.”
- Subject: Don’t Miss Out on the Email Conference of the Year!
- Subject: Join Us at the {Conference Name} Email Conference!
- Subject: Join the {Conference Name} Email Conference!
- Subject: Join us at the {conference name} Email Conference!
- Subject: Join Us for the {Conference Name} Email Conference
- Subject: Don’t Miss Out On Our Email Conference!
Rytr also lets you choose the ‘tone’ of your subject line. I chose inspirational and then informative. The two below are in an ‘urgent’ tone:
- Subject: [Name of Email Conference] – Exciting Networking and Learning Opportunities
- Subject: Network and Grow at the {Conference Name} Email Conference
(Yes, it was weak on urgency but I only gave it one shot)
All eight could use work but were written 12 seconds after my AI prompts! So if I had 20 to choose from, I could find at least two good subject lines then tweak those a bit and be ready for an A/B email test.
BTW – those above, I would edit out exclamation marks and also insert dates of the event.
Next item to test…
Email Timing
Testing the best time to send your email campaigns is smart. You can’t trust studies on the best time to open. Here’s why. Your audience is unique, so only time will tell. Unless you use triggered emails.
Unique triggered emails are even better. That’s how Inbox Mailers produces incredible open rates→ 50% – 70%. Our triggers are based on ‘inbox listening’ and the Network Effect of our partners.
We tested billions of email sends to find out this truth about email. Most people behave the same. They open their email program, delete unwanted emails, then begin opening emails they kept. Inbox Mailers triggers our customers’ campaigns to land beside those ‘keeper’ emails, boosting the odds their emails get opened too!
You should A/B test your best send times. But you should also test Inbox Mailers with a Free Demo and Q&A to see the power of perfect timing via our Network Effect.
Preview Text
Not as crucial to getting your campaigns opened as subject lines but preview text matters. It gives readers a peek at what’s inside the body of your email. Hey, if you don’t think preview text will impact your email marketing, then A/B testing can prove you correct! So test to see if preview text is worth spending time on in future emails.
Just remember to test regularly. Why? People’s habits change. Technology changes. For instance, a new iPhone could impact how your emails are viewed and previewed. You will want to know if preview text begins to matter more… or less.
And yes, AI prompts can produce your preview text as well.
From Name
Pathwire ranks recognizing the sender as the #1 reason people open emails (timing is #2). It’s crucial that your campaigns are recognized by subscribers. Especially new ones who are not as familiar with your logos, branded colors, or message tone.
Your email service platform (i.e. Mailchimp, AWeber, etc.) has settings to insert the ‘From Name’. Keep this consistent, though you may use more than one From Name for various types of emails. Example: Shawn@example.com for newsletters and info@example.com for shipping updates or invoices.
Test the From Name carefully. For new brands, choose the name with a long-range approach. If you choose a person’s name, what happens if they leave the organization? So split testing a brand versus an individual From Name is a possibility.
Have you recently changed your From Name? Check your ESP metrics to see how it impacted opens, clicks, and spam complaints.
Two ways to be more recognizable in a crowded inbox.
- BIMI is an authentication service with visual impact (your logo)
- Consider a custom branded emoji that stands out in every subject line
Text-only Versus Adding Images
A/B email testing text-only versus messages with images is easy. And not every audience will produce the same results.
Your subscribers could desire more images if the visuals tell a better story or present a clearer idea of what you are offering. Yet, another brand’s subscribers may only need text to scan through to get the info they want.
This is an easy A/B test for email campaigns. Just split test subtle changes to begin. You don’t want to shock your readers with giant changes all at once.
*Always include a text-only version of your emails (some inbox providers require this to avoid being spam filtered).
What about AI help here? Canva is in the early stages of its AI Magic Write and AI art generator. I tested some AI prompts for a fictional boat dealership using Canva’s photo library and AI for description, shown below.
*I tried an AI prompt to get Canva’s magic writer to describe the photo without any text help from me. That failed to work.
The next A/B testing opportunity is often debated.
Personalization
Read five articles on email marketing and you’ll get five opinions on personalization. Maybe even five different studies that back up each opinion!
Some mailers find that using a subscriber’s name produces more opens. Others have found the same but with a twist, it produced fewer conversions.
This tells me that split testing is the only way you can know for sure if personalized emails will work best for your brand. Here are a few options to test regarding personalization (only test two variables at a time):
- Use subscriber name in email body
- Name in subject line
- Mention subscriber’s city / state
- Show photo of an item abandoned in subscriber’s shopping cart
- Promotional email based on the specific lead magnet that brought the reader onto your list
With personalization, an AI writer can give you numerous options for the body of the email. Three AI prompt examples could be:
- Write me two sentences about Seattle’s yearly weather patterns
- Give me five variations of calls-to-action for ‘Susan’
- Write ten fun lines about what will happen to Cory’s items in his cart if he doesn’t check out at this link
See what the last AI prompt produced below. Not bad for a bot.
Email Length(s)
Why write a 500-word email when 200 words get more conversions? A/B email testing will save you some writing perhaps. Although some famous French guy or Mark Twain once wrote…
“Sorry for the long letter, I did not have time to write a shorter one.”
Point is, testing lengths will tell you a lot about your readers.
- Test shorter subject lines
- Longer paragraphs versus one-line paragraphs
- A/B test total word counts on two emails (look at past campaigns too)
Did you know some AI prompts allow an AI writer to shorten or lengthen your email copy?
Possible AI prompts for length split testing:
- Cut this email down by 100 words
- Extend the last paragraph by two sentences and add a CTA
- Summarize this entire feature story in 30 words or less
Speaking of calls-to-action…
CTA A/B Testing
Sales and conversions. Where A/B email testing can make a huge splash.
Calls-to-action prompt your readers to take the next step you want them to take. Naturally, some CTAs work better than others and can be the difference between a thin and thick profit margin!
Studies have shown buttons produce more clicks than text links. Split test this to see if you find the same results in your email marketing.
Test power words in the copy inside the button versus more subtle nudges to click the button.
CTA testing could be the most sensitive testing topic on this entire list. Why? Again, user behavior changes. Trends and technology affect this. Never stop testing your emails’ calls-to-action. Even a 2% boost in conversions will keep the competition from grabbing that 2% that should be yours.
*CTAs are usually short. Use creative AI prompts to produce a batch of them you can edit, test, and save for future use.
Email Format A/B Tests
Formats? Yes, think outside the box.
Below are two articles from Inbox Hacking (free newsletter Inbox Mailers sponsors) to get your mind going on formats:
There is nothing wrong with sending plain text emails. Those are easiest to produce, especially with AI assistance now.
But if you find you get more opens and clicks using… say infographics in campaigns, then that is vital information you can’t ignore.
Keep up with format trends to know when and what to A/B test.
Key: Go in with a hypothesis, but don’t attach yourself to the outcome you want. It’s about what your subscribers want.
Cross-selling and Upselling
Many email marketers don’t do any cross-selling or upselling. Much less A/B testing it in email campaigns.
This leaves money on the table, or in the inbox in this case! Split test to find out if customers already spending money with you are willing to buy a bit more while they are at it. I won’t bore you with proven stats on how much easier it is to sell to current customers than to new ones. You get it.
As for AI marketing prompts, this is the perfect time to use them in an A/B email test for upselling or cross-selling. See my quick example I did below with Rytr (use your bot writer of choice).
Transactional Emails A/B Testing
Don’t forget to split test upsells in your transactional emails (i.e. delivery updates) because these type emails usually get great open rates.
Your A/B testing may show subscribers have no interest in buying more even when tempted at the bottom of an invoice email. But you can’t know for sure until you run a test.
By the way, different ESPs provide their own options for testing. So if you’ve not chosen an ESP yet, take testing features into consideration.
And if A/B testing is a bit scary, you can look back at your open rates on transactional emails. If the open rates are above your average, there is opportunity there for upsells. You can simply test it for 30 days on all your transactional emails to see the impact, instead of digging into your ESP A/B testing options.
*AI prompts could work in this area but a human touch is more likely to connect with buyers.
Human-written vs. AI-written Emails
It’s doubtful you can turn an entire email over to ChatGPT. But even this could be A/B tested in email campaigns.
Would it hurt your feelings if a bot outperformed your own writing? Mine too! But I’d love to get better email engagement regardless of the writer.
Again, you can test various aspects of your email.
- Does the bot craft better subject lines?
- Are you craftier with CTAs?
- Are you better at concise copy or is the AI writer?
The split testing options are countless when you pit human copy against AI copy. Just know that all testing takes time. So don’t get overwhelmed with the process. Keep it simple and clean so you can correctly interpret the split test results.
The good news? If your AI aspects get good engagement, you can up the AI usage and save time in the long run (for more testing maybe).
A/B Testing, AI Prompt Wrap-up
Now you have options and tactics to start testing your email campaigns to find the very best versions for engagement. Is split testing painless and fast? I wish.
But it is one of the most effective ways to ramp up your email marketing strategy. Plus, the fact that testing is not painless tells you that your competition is likely not doing much testing. So take advantage of this opportunity to outperform them— level up your open rates, clicks, and conversions.
If I had to choose two testing avenues right off the bat, I’d advise testing subject lines first. Second, I’d test timing with Inbox Mailers because this email tool does most of the work for you when it comes to getting your emails noticed and opened.