So many brands rely on email campaigns to do business directly with consumers who sign up to hear from them. However, mailbox providers have a huge influence on how effective your email marketing is.
Your list may be composed of your customers and leads. But those customers and leads are also a mailbox provider’s users.
That means mailbox providers serve the user, not you, first. And over the last two decades, the number of those users have exploded. In 1995, there were 1 million email users. In 2015, there were 2.5 billion.
To cope, mailbox providers now rely on sophisticated artificial intelligence (AI) systems to serve users. And these AI systems have a big effect on you.
They determine if you’re spam or not. They decide how trustworthy you are. And they even categorize your messages for users, telling them what type of communication you’ve sent.
In short, AI has a say in how effective all your email marketing campaigns are.
That means marketers must better understand how AI in email marketing impacts their efforts to communicate with their list. Armed with this knowledge, marketers can then build smarter campaigns that result in better metrics.
To start doing that, it’s important to understand the two big ways that AI affects your email marketing:
Deliverability and engagement.
Artificial Intelligence and Email Deliverability
Email spam filters are powered by AI. But these filters have evolved since the early days of simply marking an email “spam” or “not spam.”
Today’s spam machine learning models take into account thousands of factors and user behaviors to predict which emails are garbage and which aren’t.
This is an imperfect science. Sometimes, you still see important emails go to spam. But that should be a lesson to all marketers. There are still many ways in which honest, valid communications can run afoul of the rules that AI systems use.
The problem is: We don’t know most of these rules. The AI systems that power mailbox providers are proprietary. Providers aren’t giving away the secrets to how they operate any time soon. And AI systems are continually tweaking and improving the rules in real-time. This would make it difficult, if not impossible, for any human to follow all the changes and logic that affect the outcomes we see in inboxes.
It’s very similar to Google’s AI-powered search. We know some best practices. We can closely watch what updates the company makes, so we can adapt to them. But we have no fully documented list of the rules this highly complex AI uses to make its decisions.
The same goes for the AI used by mailbox providers (including Google). Providers may offer guidelines that marketers can follow to increase the chances of deliverability—and these are good rules to follow. But the full list of factors that AI uses to classify your email campaigns is not public knowledge.
This matters more than you might realize.
One marketer for a major enterprise we met at Return Path told us about the extremely low open rates in her campaigns. We pulled up some data on her company and saw that only 2% of the emails were making it past spam filters.
Before you adjust anything else—your spend, your strategy, your execution—you must solve for deliverability.
This is easier said than done. In the case above, recognizing the problem is a great start. But it can take a lot of time and effort to uncover the root causes of deliverability issues.
If you don’t uncover the problem and fix it, you could literally be sending millions of emails that nobody clicks because they aren’t seeing those emails in the first place.
Artificial Intelligence and Email Engagement
AI systems used by mailbox providers don’t just determine email deliverability. They have a lot of power over how users engage with your emails, too.
Take Gmail as an example.
When you send an email, Gmail’s AI sorts it into different categories, or “inboxes”, based on its content and past user behavior. Once Gmail determines your email isn’t spam, it then sorts you into the Primary inbox or the Social, Promotions, Updates and Forums inboxes.
As a brand, you may not have much control over where Gmail or another provider sorts your email initially, especially if you’re sending promotional emails. But how the provider sorts your emails can change over time, as the AI adapts to signals from user behavior.
If a user consistently puts your communication in the Primary inbox, Gmail’s AI will adapt to sort future communications from you differently.
Gmail is not the only system that adapts based on user behavior. Other email clients make decisions based on how users open, click, sort, and engage with your email communications.
Some marketing automation platforms used to send email, like HubSpot, will even opt out of sending unengaged contacts an email. This is beneficial for brands, but, like the systems used by providers, is entirely driven by user behavior.
In short, AI enables almost total user control over the effectiveness of your email marketing campaigns.
Marketers can no longer barge into someone’s inbox, then hook them with a killer subject line. You must instead plan a longer-term email strategy that focuses on creating immense value for email users.
That means you must send the emails they want to receive, not the ones you want to send.
What Should Marketers Do Next?
Alright, so AI has way more control over deliverability and engagement than you realized.
That’s great, but what do you do about it?
At Return Path, we help businesses communicate more reliably, effectively and more securely. We help all types of businesses make sure their emails get delivered and engaged with, in the way they want, producing the results they need.
And in most cases, we’ve found the best first step is to use human experts or machine tools to gauge your current deliverability and engagement rates. You’ll also want to determine your sender reputation, which informs how AI models treat your emails.
To do that, we recommend you take our free comprehensive reputation measurement assessment here.
Lauren McCombs
Lauren McCombs has been at Return Path for 3.5 years and currently manages the Data Science team.