Most Common AI Email Marketing Mistakes

1 April 2021Ashley Maxwell10 min read

We all know the classic email marketing mistakes, but these days, thanks to AI machine learning, people are making email marketing mistakes that are spurred on by AI automated tools. In an odd way, AI tools have made many marketers more efficient at doing a bad job. Here are some of the most common AI email marketing mistakes.

Not Giving the Reader Enough Information

Let’s say that you have set up an email list. The people on your list are all there of their own free will, and they have completed a double-opt-in procedure (confirmed opt-in), so you are sure they wish to receive your marketing message. You use your AI marketing to select the right candidates, and you send marketing messages based on their preferences. In the strictest definition, they are unsolicited marketing messages, but they are not undesirable emails, and they are not being hoisted on people who do not want them.

Write this on your wall above your desk, “Do not try to entice your email recipient”. What are known as “Baited email subject lines” are severely damaging to your online reputation. Years and years of spam email marketing have hardened even the dumbest of a recipient. When you try to entice your recipient by offering a bit of information, but not enough information (baiting them), it doesn’t compel them to click your link or follow your offer. It compels them to block your message.

The Problem – Baiting People Does Not Work

The classic college response is that you need to generate interest, and enticing the reader with a snippet of information is a way to do that. But, the classic college response ignores the market conditions. It ignores the fact that your recipients, even kids, have been hardened to this sort of marketing. Trying to stimulate interest by offering a snippet of information with the hopes of getting people interested is a fool’s game. It is not the way to entice a modern email recipient.

The most enticing email adverts are the ones that are upfront, honest, and that offer genuine value to the recipient.

If you are old enough to remember “The Muppets Take Manhattan” (1984), you may remember them poking fun at what they considered to be modern marketing methods. Kermit was taking part in the brainstorming group where the other executives were offering overly long or buzzword-full suggestions for advertising slogans. Kermit comes out with the line, “Ocean Breeze Soap – It Gets You Clean.” The team of executives lose their mind at innovative his thinking is.

The line was a jab at how marketing professionals blur the issue, and how they are not direct. Yet, even today, people are trying to sell the sizzle when they have a perfectly good steak that people want. If your products do not offer genuine value to your customer, and if you are having to always go for the hard sell, then perhaps the problem does not lie with your marketing (AI or otherwise).

Advice – Sell the Steak – Not The Sizzle!

These days, your emails should be honest and very clear. It may seem like you are showing your hand a little too early if you give the offer upfront without leading them to your landing page, but the fact is you need to appeal honestly to the recipient right from the start. You do not have the luxury of assuming your landing page is going to make the sale. The lead firstly needs to be warmed through the content of the email. Then, if they decide they are interested or want to know more, then they find their way to your landing page where your on-page marketing does the rest.

Auto-Response Emails That Offer Up Too Much Information

It is very important that you give your email recipients all the information they need to make an informed buying decision. As mentioned in the section above, you will get nowhere trying to entice the user by holding back too much information. However, the rule is not the same with auto-response emails to customer enquirers. Let’s say that you are selling robotic lawnmowers, and you receive inquiries via your website. People receive two auto-generated responses when they contact you, and both are used to convince the recipient to buy your device. The first answers their direct questions, and the second is more of a marketing message.

The Problem – Auto-Response Emails Should Be Closing, Not Selling

The trouble is that the marketing emails are too broad. If the recipient is making enquirers, then the recipient is already part way to making a purchase. It is your job, and the job of your AI marketing tools, to figure out what the final objections are so that they may be overcome.

However, what often happens is that the objections are/aren’t answered, and the marketing messages create more objections. For example, they may start boasting about how the device can go for 1 hour without needing emptying, which then reminds the recipient of the hassle of changing the clippings bag. Or, the advert may brag about battery life, when then reminds the user of the hassle of having to recharge when the job is half done.

Would you like to hear my life story

There is a time to talk and a time to “Shhhh.” If your auto-response emails are not addressing the problem in hand, if they are trying to sell, then they are not doing their job. Remember your ABCs (Always Be Closing). If your AI tools are offering up more promotional information when they should be answering questions, then they are resetting the sales process–not pushing it forwards towards a close.

Advice – Optimise Your Auto-Responses a Little Better

When it comes to auto-response messages, your job is not to overtly market your product. Your job is to figure out which objections the potential customer has, and to overcome said objections. You can optimise your emails pretty easily with things as simple as “Select the problem” contact forms. If your AI tools have at least a small idea of what the problem is, you can craft suitable automatic responses. Just don’t stuff your responses with too much information, especially promotional information.

Focusing on Opening Rates

We all want our emails to be opened. In the old days, the only way you knew if your emails were opened is if people followed a custom link that had a unique identifier on it. Luckily for marketers, these days, emails have read receipts and email tracking. Though many savvy users turn them off on their email accounts, and some email accounts do not offer read receipts. Nevertheless, it is possible to get a vague idea of how many people are opening your emails, even if the number you see is often lower than the actual number.

The Problem – You Are The Problem

AI marketers focus on email opening numbers way too much. Most people see your email subject line and a part of your email before they open it. AI marketers think that if people are not opening their emails, then something is wrong, and the subject lines need to be changed, or the offers need to be changed. They end up changing offers, changing headers, and perpetually honing emails that may have already been perfect.

The fact is that sometimes, the perfect email is perfect because it wasn’t opened.

The notion that every one of your emails must be opened is childish. It also grossly misunderstands the marketing process as a whole. Perhaps in the future, if targeted marketing means mind reading, then your open rate will start to matter.

Advice – When You Do Something Right…

Your AI may have picked the perfect person based on their purchase history, their demographics, their previous interactions, and so forth. You craft the perfect email subject line with the perfect offer inside, and the email isn’t opened. To the novice or the reactionary, this is bad news. To the seasoned professional, this is neither good nor bad news.

When you do things right

The key to your success is that people who want your message/offer will open it, and people who do not want your message/offer will not open it. It may sound like an oversimplification, but it is very important. You may have done everything right, but it just so happens that the user doesn’t need your offer right now; he or she may have received your product already as a birthday present. If somebody who doesn’t need your offer, you do not want them reading your email. If you are convincing people who don’t want your offer to open your email, you are wasting their time, which puts you one step closer to getting blocked or them unsubscribing.

Focusing Too Hard on Avoiding Spam Filters

When you send automated messages, especially on bulk, you are more likely to be caught in the spam filter. There are a series of things you can do to avoid being caught in the spam filter, but learning them is pretty pointless because they are always changing. Spammers are always adapting to cheat the system, and so spam filters are always adapting.

The Problem – You Care Too Much

Some companies cannot accept the fact that their automated emails are repeatedly being pushed into the junk folder, and they work hard to make sure they remain in the inbox. They use the 20/80 ratio of pictures to text, they run campaigns off rotating servers and IP addresses, they send basic text messages, and so forth. Their creative standards and marketing standards slipped as they tried their best to fit their designs into the cookie-cutter shapes that were the non-spam red flags.

Advice – Spam Filters Are Not Your Enemy

Your recipients have double opted in, they want your emails. If they end up in their junk file, they will find them and pull them out. Plenty of people check their junk files and pull things out, it is a classic procrastinator’s time waste, especially in business environments where access to social media is blocked. Plus, some people purposefully keep things in their junk files. It gives them some easy reading whenever they are done with their important email stuff. Stop worrying about being caught by the spam filters. Ending up in the junk folder is not as bad as you are being led to believe.

Sidenote: Being marked as “Junk” and being “Blacklisted” are very different things. Having your keywords, company name, IP address or servers blacklisted is something you want to avoid. Here is how people get blacklisted, these are the things you need to avoid:

  • You sign up with an emailing company that has banned servers because they do not guard against spammers.
  • You are reported too many times for sending Phishing emails.
  • Your email address is hacked and used to spam people
  • You overtly break the law with your emails

We at are not going to recommend which bulk-email-sending companies you can trust. But, if an email-sending company makes it difficult for you to spam, and if they insist on double-opt in email lists that you build yourself with their tools, then the email-sending company is probably protecting its servers and IP addresses very well.

Missing on Market Research Because of Segmenting

You want your AI email marketing software to expertly target your potential consumers. It is a learning machine, so it is helping you figure out who buys what, why, where they are, and how they interact with your emails and your website. The trouble is that when you narrow down and pigeonhole your recipients in this way, you lose the ability to learn more from them.

The Problem – People Confuse Segmenting With Targeting

So good was the marketing executive at segmenting, that she had offers perfectly targeted to each user based on their previous purchases. Yet, this method only ever sparked repeat sales over a long period of time. When a customer bought a lawnmower, then that customer received a lawnmower and lawn care offers for five years before ever deciding to buy another lawnmower. Yet, that same buyer has kids, an ex-wife, two Yorkshire terriers, and is looking to buy a pool. Even if the executive was just selling garden equipment, how about sending the occasional email about kid’s garden toys, pool cleaning robots, dog-proof fences and so forth.

Advice – Forget Segmenting From Time to Time

One assumes you are spending at least a portion of your time on market research, along with doing a little A/B testing into the mix. It is important that you ignore segmenting now and again and send something a little different to your users. Even with the best information gathering on the planet, you can never understand the full picture.

  • Does the manly and gruff biker buy Barbies? What if he has a young daughter?
  • Does the young single mother buy extreme snowboards? What if she is a snowboarding champion?
  • Does the elderly grandmother need hot-rod alloy wheels? What if she shares the advert with her toy-boy lover who visits twice per week?

Segmenting is fine, but pigeonholing people all the time is dangerous. Mix your market testing with a few randomised email offers to see what else you can learn about your recipients.

Not Testing Your AI-Powered Personality

You have probably read lots of articles giving off terrible advice about AI-powered email marketing. In fact, I wrote an article debunking bad online AI email marketing advice. You may have seen articles where they say things like, “Add personality to your emails.” It is dumb advice because it often prompts people to add social media speak (hashtags LOLLING), or to add jokes (as the confused camel said to the chair, “Moo”), or they try to over-personalise (Hey, George Smith, it’s been a while, how ya bin?). The truth is that you probably shouldn’t add personality to your emails because people are aware they are being sent by massive server rooms in other countries. However, you should strive to make your emails a little less clinical and sterile.

The Problem – Personalisation Does Not Equate To Personality

An AI programming consultant showed a company how to use the find-and-replace functions on the AI email senders to make emails more personalised. Portions of the email were coded so that they changed for each user. The first was the name on the email, where a placeholder was substituted for the user’s name. Other elements included the user’s last purchase, time since last purchase, the user’s location, and the user’s profile picture next to a sign that says, “Our best customer.”

Dave, the chartered accountant receives an email to “Doris Backwater,” the fake name he used to sign up. It shows his last purchase, the nail varnish his daughter bought on his account. It showed the time he purchased last, which was 3 years ago (not very impressive), and it showed his location, which Dave found creepy. Finally, it showed an empty avatar next to a sign saying “Our Best customer” because Dave never uploaded an image to his linked Gravatar or Instagram account, which is where the AI marketing program gets its avatars from.

To First Name

Advice – Personalisation Should Have a Direct and Convertible Goal

Personalisation is not window dressing, nor is it a clever marketing tactic. Personalisation is a means to an end, and it should be done with extreme subtlety. When you are personalising, unless you are pulling names from payment information records, do not address people by their names in emails. It doesn’t make things more personalised, and it is a ripe area for getting it wrong. Do not draw attention to what people bought last time, and instead, auto-generate adverts for related products, like if Dave had bought a printer, you could show your compatible printer ink offers. By all means, use your AI marketing tools to create emails that are better targeted, but do not confuse targeting with personalisation. People would prefer engaging emails, rather than emails that predict their next pregnancy, or that comment on driving tickets, etc.