How AI is Used in Email Marketing

25 January 2021Ashley Maxwell8 min read

These days, when people talk about AI, they do not mean a living machine. They mean a sophisticated decision-making piece of software. The biggest steps forward in AI have all come from the gaming industry. Yet, we shouldn’t discount the online marketing industry for its great strides in the area of AI marketing.

Not only has the process become a lot cleaner and targeted, but the “Use” of artificial intelligence in email marketing has become astoundingly easier. Just fifteen years ago, if you wanted a perfectly optimised email-marketing system, you needed a programmer, analytic expert, and whatever passed for a big-data processor. Nowadays, you can sign up with a marketing platform and have your email marketing set up within a day, and optimised within three to six months (being careful to avoid these common mistakes). Here are just a few ways that AI is utilised in email marketing.

Brand Building Through Drip-Marketing

We all want people to be exposed to our brand in a positive and/or educational way, but thanks to years of spam emails, it doesn’t take much for a dedicated customer to re-direct your emails to their spam folder, or to block your emails completely. Drip-marketing works through repeatedly exposing people to your brand without overly pushing a sales narrative. In most cases, it is less about getting people to open your emails and more about stopping people from blocking you.

Email marketing AI helps avoid any form of email saturation by timing email frequency based on user engagement. Let’s say you are a FOREX trader, and you receive both exchange rate updates and promotional offers from a trading platform. If you frequently click the links in the emails, then the AI sends emails more often (perhaps even daily). On the other hand, if you rarely engage with the emails, and rarely log into your account, then the AI sends fewer emails, spacing them out a little more, and only sends offers and reminders rather than daily update information. If you start engaging a little more, perhaps through logging into your account more, then the AI starts increasing your email frequency.

AI-powered brand building is not limited to timing and engagement triggers. It uses trends to adjust the content and frequency of marketing emails. For example, in the run-up to the highly anticipated season four of Rick and Morty (a popular TV show), online media companies increased their frequency of emails as the TV show’s airing date approached. Consumers were sent everything from animated clips, behind-the-scenes footage, and fan theories. As the show’s air date approached, trends indicated a higher degree of interest, so more emails were sent out, and more opportunities for branding were presented.

Cutting Customer Support Costs

This is perhaps one of the most readily-available and easy-to-use email marketing functions of the modern era. With the use of AI auto-responses, you can drastically reduce the amount of human interaction needed when you offer customer support. Plus, it is an area where you really see the difference when you optimise your system because the demands on your human support staff go down almost immediately after your optimisations take effect.

For example, let’s say that you sell a PC game, but a recent Windows/Mac update means that gamers now need a driver update. On your contact form, you add the option, “My game does not work after an operating system update.” When the AI email system receives emails from somebody who selected that function, it replies giving several fixes, including a link to the new driver and any related installation instructions. It also gives the user the option to escalate the problem and receive a tech support ticket if the emailed fixes do not work.

There is also a fair amount of groundwork that can be laid through AI email auto-responses. In the example above, if somebody does decide your emailed fixes do not work, and they receive a ticket for tech support. Then, when the tech support staff member is reached, he or she already knows the problem. They know what has and hasn’t been tried, all of which saves a lot of time for your human tech support staff.

What’s more, is that the user doesn’t know that the emails are automated. Depending upon the type of queries being posed, it is quite possible for the email responses to appear as if a human wrote them.

Clever Suggested Product Offers

You are probably familiar with email AI that uses your purchase history and wish list to guess at which offers it should send you. For example, if you bought car parts and bike parts from an online superstore, they send you automobile-related offers rather than kid’s toys offers.

The use of your wish list is a common feature for PC and Mac gamers. You put things on your wish list, and you receive emails when those items are on sale. You also receive offers when new DLC for the games come out.

Suggested product offers have become more sophisticated thanks to big data. AI marketing programs can guess what you want when you want it, and which offers will tempt you into buying. They can do it by using your purchase history, which includes using data from loyalty schemes such as club cards. In fact, in one example the store “Target” was so good at interpreting the data it pulls from customer’s accounts that it was able to guess a teen girl was pregnant before she did.

Generating Sales Leads

Using the mechanics mentioned already, an AI-controlled email marketing system can generate sales leads. Sometimes, they are simple sales leads that arise from sending offers based on purchase histories or wish lists. Sometimes, they are sales leads based on a back-and-forth that heats up a lead before a real person is contacted. Other times, the use of drip-marketing, and the ability to get people to engage with your emails (or at least open them) is enough so that when they do decide to buy, then yours is the first company they think of.

How you optimise your AI-controlled email marketing depends on how you run your business, your target audience, and the reasons why people signed up to your email mailing list. The great thing is that these days, you have access to a wealth of demographic information, along with associated analytics, which means things like location-based targeted can be allied with your own sales information to better hone your email marketing.

If you were building a sales funnel, you can use AI email marketing at almost any juncture in order to generate sales leads. Here are a few sales funnel junctures and reasons why AI email marketing may help:

  • Building interest – Offers based on trends or promotional directives.
  • Cold leads – Email offers based on demographics.
  • Information Gathering – Emails prompting the user to offer information.
  • Cold offers – Emails based on account activity.
  • Leads – Email offers based on purchase history and user activity.
  • Offering solutions – Emails based on previous inquiries and search activity.
  • Trending solutions – Emails based on demographic trends (e.g. offering fans in summer).
  • Education – Emails offering product information, examples, demonstrations.
  • Targeted Offers – Emails based on user interaction with the website’s sales pages.
  • Email Triggered Offers – Emails based on how the user interacted with the emails.
  • Cart Retention – Email people who have not committed to buying just yet.
  • Buying Upsells – Emails acting to re-enforce your current up-sell strategy.
  • Cart Abandonment – Emails to remind people what they left in the cart and didn’t buy.
  • After-sale – Email a receipt with an exclusive discount offer attached.
  • Aftercare sales – Email offers for replacement parts, refills, etc.
  • Word-of-mouth marketing – Emails with discounts if you sign up your friends.

Note that building awareness is mostly missing from this funnel because if you are following the data protection laws in the UK and in other jurisdictions, then your recipient has agreed to receive your newsletter, which indicates they already know who you are.

Every marketing email example above could be sent by a clever AI marketing system. Just like a dynamic website, artificial intelligence could use what it knows about the potential customer to better target the individual.

How difficult the process becomes will vary depending upon the sales funnel juncture. For example, optimising your AI email marketing for cart abandonment is easy. You simply have it wait a few days, and then it sends offers and variants of whatever items were left in the cart as it was abandoned. It doesn’t take much optimising to get your email marketing system to do that. Something like targeted offer emails based on a user’s interaction with a website is trickier, especially if the user isn’t signed into his or her account while exploring your website.

Reminding Subscribers to Engage

The UK government and the 2016-2020 Trump administration brought in measures to stop the exploitation of vulnerable people through trigger marketing, also known as black-hat emotional marketing. We are all aware of smoking triggers, in that reminding ex-smokers of smoking is a great way to win them back. However, it is now more difficult for gambling and alcohol marketers to prompt (trigger) a recipient’s addiction through advertising.

The new measures have extended far beyond email marketing into areas like reality-checks on gambling websites, stronger watershed advertising rules on TV, and banning online adverts on websites where vulnerable people (addicts, the young, the mentally ill) may see them.

Supposing you are not using email marketing for nefarious means, there are ways you can remind users to engage with your service or product through clever email marketing. Have you ever wondered why takeaway emails arrive over the weekend? It is because AI email systems examine broad trends, local demographics, and your online activity and decide that the weekend is when you are most likely to buy a takeaway.

Gambling websites in the UK use similar (legal) methods. For example, if you do not deposit into the gambling account within three months, you often receive an offer via email. Sometimes, as with the case of Bet365, you actually receive free gaming credit if you were an active user, but you haven’t deposited in a while. The decision as to when to send the emails, and which offers to send, is not made by humans. It is made by artificial intelligence.

Customer-Specific Service Emails

There are a plethora of different ways that customer-specific service emails are used in modern email marketing. The best way to explain them is with examples.

Take BMW for both people who buy from their showrooms, and for people who take their BMWs to be serviced or repaired at their official locations. Not only do you receive emails telling you when your next car service is, or when the next MOT check is, but they also use the age of your car and your purchase history to guess when you are next likely to start shopping for a new car.

If you were taking a Japanese language course, and you received a “Word of the Day” email from the company. They may use your account status, your learning status, and any online achievements within the website to decide which words to send you. Their up-sells are also keyed into learning and achievement status. JapanesePod101 take it a step further because when you have completed most of their higher-level courses, they start affiliate-selling trips to Japan both on their website and their word-of-the-day emails.

Free-to-play MMOs often send you in-game purchase offers based on your current gaming level. If you are signed up with Ubisoft directly, and you receive their newsletter emails, then when you level on up games like Assassin’s Creed, they send you in-game purchases that are relevant to your level. E.g. they do not send offers for 100 gold coins if you already have 50,000 in your account.

Final Thoughts – Do You Need To Fully Understand the Mechanics of Email AI to Use It Correctly?

No, you do not. You do not need a degree in programming, you do not need a mathematics degree or any experience with big data processing. All you need is a willingness to learn, and the simple understanding of trial-and-error testing.

One of the great things about email marketing platforms is that they allow you to start out in a very simple way. As you grow more accustomed to the rules and grow to learn how to use your own analytical data to hone your email marketing, you may then customise and optimise your email AI to become more effective.

Do you remember how interacting with pre-90s computers used to be difficult, requiring you to remember hundreds of terminal commands? Now with graphical user interfaces, most people can use a computer without prior knowledge.  That is how email marketing AI has become. Where you used to need a degree in programming, you now only need a quick video tutorial on how to use the email marketing interface in order to get things up and running. Then, after a lot of trial-and-error testing, you can have a well-optimised AI email campaign within three to six months.