Marketing AI Institute | Blog

How to Do Better Email Marketing with Artificial Intelligence

Written by Mike Kaput | Aug 29, 2019 2:01:00 PM

Artificial intelligence gives marketers the ability to increase revenue and reduce costs at an unprecedented scale. And AI in email marketing is one of the top ways to boost your performance.

In fact, AI is already being used to write subject lines better than humans at scale and optimize send times.

But how do you actually get started with AI for email marketing?

This post is here to help.

It will walk you through actionable steps to identify email marketing use cases and pilot AI solutions for each one.

How to Think About AI Use Cases in Email Marketing

When coming up with AI use cases, we find it useful to have a framework to organize potential applications for AI.

That's why we created the 5Ps of marketing AI framework. The framework organizes use cases into five broad categories that relate to marketing activities:

  1. Planning — Building intelligent strategies. Sample use cases include: determining goals; constructing buyer personas; segmenting contact databases.
  2. Production — Creating intelligent content. Sample use cases include: writing nurturing emails; converting voice to text; designing website.
  3. Personalization — Powering intelligent consumer experiences. Sample use cases include: recommending content; optimizing send times; product recommendations.
  4. Promotion — Manage intelligent cross-channel and cross-device promotions. Sample use cases include: scheduling social shares; delivering ads; testing headlines.
  5. Performance — Turn data into intelligence. Sample use cases include: lead scoring; forecasting performance; writing performance reports.

 

Below, we'll use a few simple strategies to brainstorm use cases, then use the 5Ps framework to organize and prioritize which use cases are most important to investigate further.

How to Identify AI Use Cases in Email Marketing

1. Make a list of email marketing tasks you do on a daily, weekly, monthly, and quarterly basis.

To start, make a list of all the tasks related to email marketing that you do on a daily, weekly, and monthly basis. These can be as simple as "uploading email copy" to complex tasks like "create email marketing strategy." Sample tasks could include:

  • Write email subject lines
  • Upload email copy
  • Create email marketing strategy
  • Build workflows
  • Create email newsletters
  • Experiment with email send times

 

Next to each task, list out "daily," "weekly," "monthly" or "quarterly" next to each task depending on how frequently you do each one.

Each of these tasks represents a potential use case for marketing AI.

Note: Don't worry right now about whether or not AI can do the task/use case you list. Just write down what you actually do in a day, either manually or using 

2. Organize each task/use case within the 5Ps framework.

Now, do your best to assign each task/use case to one or more of the 5Ps. This will help you both organize the items you've brainstormed and inspire other potential ideas.

Here's a sample of what an email marketing 5Ps use case list could look like.

  • Planning
    • Plan content for newsletter - Weekly
    • Create email marketing strategy - Quarterly
    • Create newsletter strategy - Quarterly
  • Production
    • Write marketing email copy - Daily
    • Upload email copy - Daily
    • Create newsletters - Weekly
  • Personalization
    • Write custom subject lines based on database segmentation - Daily
    • Customize newsletter based on database segmentation - Weekly
    • Build workflows based on database segmentation - Monthly
  • Promotion
    • Experiment with email send times - Daily
    • Attract new subscribers - Daily
    • Ensure email deliverability - Weekly
  • Performance
    • Analyze open rates - Weekly
    • Analyze click rates - Weekly
    • Produce executive performance report - Monthly

3. Assign a "Value to Intelligently Automate" to each task/use case.

We use a metric we call "Value to Intelligently Automate (VIA)" to assess AI use cases. The VIA is a subjective rating you give every use case from 1-5, with 1 being "no value" and 5 being "transformative."

When considering how valuable a use case would be to automate, take into account the following:

  • How important the task is to business/marketing goals
  • How often you do the task
  • How many people are required to do the task
  • How long the task takes

 

Your most time-intensive tasks or tasks that require many team members are often ones that end up with a high VIA.

4. Pick the highest-rated task/use case to research and pilot.

Once you've assigned the VIA to each task/use case, it's time to identify which one to research and pilot. 

During the research and pilot phase, you'll be looking into whether or not AI can actually do the task and, if so, what vendors/tools to begin piloting.

There's no guarantee you'll be able to use AI for a specific use case. So, you'll want to start with your highest-rated use case, then move down the list if your research produces few results.

Your research can be as simple as searching for terms like "AI for [USE CASE]." You can see below some of the marketing categories we've covered so far:

 

You may also want to accelerate your research with our AI Score for Marketers tool. AI Score is a free assessment that helps marketers rate over 60+ common marketing AI use cases, then receive customized vendor recommendations. The assessment takes just a few minutes to complete.

Last, but not least:

If you're really stuck in determining use cases or identifying vendors, feel free to contact us here and we'll see if we can help.