How to Create and Test Subject Line Variants Using AI

How to Create and Test Subject Line Variants Using AI

“Wow, wait a minute” is the thought many of us have had after something caught our eye in our email inbox.

When it comes to standout emails like these, their creators have perfected one element – the subject line.

If you send cold emails at scale, you should use the opportunity to test multiple subject line variants to eventually find the perfect one.

But how to come up with variants to test without running out of ideas?

The answer is AI (shocking, right?), and in this article, I’ll show you exactly how!

Subject lines and their role in cold outreach

Cold emailing can be as intimidating as it sounds for both the sender and receiver. At the end of the day, cold emails represent initial contact with someone completely unfamiliar. To make things worse, the average person receives 100 to 120 emails a day, making it very likely that unsolicited emails from random senders will go overlooked.

The subject line is what will have your email stand out, or if done very poorly, have it end up in the trash. Backing this statement up is a study by SuperOffice stating that 33% of people use the subject line as a deciding factor when it comes to reading an email or simply disregarding it.

Why is having different tone-of-voice subject line variants important?

In writing, tone represents your word choice, sentence structure, as well as the way you order information. When put together, these three can showcase the attitude you hold towards a topic as well as the reader of your writing.

Now, when it comes to different tone-of-voice subject line variants, you can think of them as a tool to boost engagement and open rates as well as add more personalization. But, when done right, they can bring even more benefits, such as:

  • Having an easier time connecting with various segments of your audience and better resonating with them,
  • Aligning your email communication with your brand,
  • Keeping your communication fresh, interesting, and non-monotonous,
  • Emphasizing the aspect of your message most relevant to each recipient.

To help you better understand this whole concept, here are some examples of different commonly used tone-of-voice subject lines:

  • Urgent/action-oriented:

"Last chance"

"You’re missing out"

  • Curiosity/intrigue:

"I found this on your website"

"What your customer said"

  • Friendly:

"Love your new post!"

"I thought you’d love this"

  • Direct/informative:

"New product announcement"

"Join the webinar today"

  • Gratitude/appreciation:

"Thank you for being a part of this"

"We're grateful for your ongoing support"

  • Problem-solving:

"This is the solution to {{problem}}"

"Say goodbye to {{problem}}"

How to use AI to generate different tone-of-voice subject line variants

As with most AI solutions, subject line generators need you to provide input about your specific use case. Once they have the data, they can then generate effective subject lines in different tones of voice thanks to the natural language processing and machine learning algorithms they’re based on.

Let me show you how you can quickly generate various subject lines with AI. I'll use ChatGPT here.

First, this is my email body:

📧
Hi Ziemek,

I've been using Hunter for 5+ years now, and love the product. I write long-form content for B2B SaaS companies, with over 1,000 bylines from over a decade of experience. If Hunter ever needs any writing done for the blog or newsletter, I'd love to help.

Nice to meet you!

And this is the prompt I used with ChatGPT:

💡
This is a cold email. {{insert email body}}

Please help me generate different subject line variants for this message. I'm looking for subject lines that employ the following tones of voice:
enthusiastic,
direct,
restrained.
For each tone of voice, come up with two subject lines.
They can't be longer than 6 words.
To evoke interest in the recipient, the first subject line for each tone of voice variant should use the curiosity gap technique, while the second one should use controversy/provocation.
Only capitalize the first word in every sentence.

Here's what ChatGPT came up with:

ChatGPT output based on the prompt above

Do keep in mind that even though AI is very powerful, it’s still prone to making mistakes, and that is why you should carefully read each subject line it generates and edit it accordingly if necessary.

AI tool recommendations for subject line generation

When it comes to subject line generation, there is definitely not a shortage of AI tools. Having said that, you shouldn’t have a hard time finding a tool that has the features you need and fits within your budget.

If ChatGPT doesn't cut it for you, here are some other recommendations!

Yet Another Mail Merge (YAMM)

Yet Another Mail Merge (YAMM) offers a set of AI tools for drafting emails, including one for subject lines. With it, you can create great subject lines quickly and easily while also making them personalized, creative, and effective.

Necessary information:

  • Email body
  • Tone

Result example:

Results generated by YAMM subject line generator

Mailmunch

Another free and GPT-3-based email subject line generator is offered by Mailmunch. The generator going under the name of Copy Assist can generate “killer, compelling, and catchy subject lines” with just basic prompts and guarantees higher open rates for your emails.

Necessary information:

  • What is the email about
  • Company
  • Tone

Result example:

Results generated by Mailmunch subject line generator

Stripo

Stripo email template builder offers an AI content improver that generates and enhances subject lines. It leverages advanced algorithms to analyze and suggest subject line improvements to assist email marketers in crafting more compelling and effective subject lines for their campaigns.

Necessary information:

  • Business specialization
  • Prompt describing what the email is about

Result example:

Results generated by Stripo subject line generator

How to test AI-generated subject lines

Regardless of whether they are human or AI-generated, subject lines should always be tested so their effectiveness, relevance, and impact can be optimized.

Testing email subject line rendering

The first and maybe most crucial type of testing you should put your subject lines under is testing of the subject line rendering. This entails checking how the subject line looks on different email clients, devices, and screen sizes so you can be sure the rendering is consistent and appealing.

To test how your subject line will be rendered for the recipient, you can use Mailtrap.

Mailtrap Email Testing is intended for inspecting and debugging emails in environments such as staging, development, and QA without the risk of spamming recipients with testing emails. The tool comes with features for checking the support offered by popular mailbox providers for an email’s HTML and CSS, spam score analysis, blacklist reporting, and more.

For subject lines in particular, Email Testing offers a feature for previewing how the subject line looks to the recipient. The feature can be found in the Tech Info section of the Mailtrap dashboard. There, you can also check the SMTP transaction info and email headers.

Mailtrap email tester interface

A/B testing email subject lines

A/B testing is a common form of testing that entails creating multiple variations of the same copy, subject line, etc., and sending them to different segments of your audience to compare performance. The results of this type of testing can be used to optimize subject lines and get better open rates, click-through rates, and engagement.

Conclusion

Before you begin looking for the most optimal AI tool for you and crafting those subject lines (no matter the type of marketing you're going to use email for), make sure to keep in mind that:

  • Different AI tools will ask for different inputs for subject line creation, so pick the one you are most comfortable using.
  • There is a range of free tools available, so do give them all a go first.
  • Never skip out on testing (A/B or subject line rendering).

Good luck!

Was this article helpful?
Dzenana Kajtaz
Dzenana Kajtaz

I’m a Technical Content Writer at mailtrap.io with an educational background in software engineering and years of experience as a writer, editor, and content marketer.