Tiffany: Artificial Intelligence in Email Marketing

2017 was probably AI’s breakout year in email marketing. It's from then on that I started to witness the results that AI can bring to the channel primarily when used in combination with Machine Learning and Automation.

Change has been rapid and ongoing since then, and it doesn’t look like slowing down any time soon. So, what’s the current state of play? How is it affecting the way we market? And how are things shaping up for the future? Here are my thoughts on AI in email marketing:

The basics:

We’re all seasoned professionals here, but there is often a misconception about what Artificial Intelligence is so here’s a definition from Oxford Dictionary:

'The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.'

Think predictive customer services, when you know when and a why a customer is contacting your business, and for email marketing, it's the ability to automatically customize email content based on the individual recipient’s behavior.

Most people picture AI as robots taking over. AI is not here to be scary, and I'm going to share how other businesses have used AI successfully to optimize email marketing activity. It’s important to remember the facts before we fear for our jobs:

• AI Creates More Jobs Than It Takes
• By 2020, AI will become a positive net job motivator, creating 2.3M jobs while only eliminating 1.8M jobs​
• More than replacing humans entirely, AI will augment existing employment and improve productivity.​

Source: Gartner 2018

We're able to collect now a vast amount of data about subscribers (provided consent has been given), and AI enables us as marketers to harness that data. I'm a particular fan of how Gartner describes the relationship between Humans and Machines as being symbiotic, one cannot work without the other.

The key benefit of AI is that it enables us to personalize at scale.

But to personalize and tailor messages at scale, we’d need contextual data on thousands or millions of people. Then, we’d need to analyze this data and use it to create the right communications for each person.

Humans don’t have the time, resources, or intellectual bandwidth to do this. But AI systems do.

That’s why more and more companies are turning to AI as the solution to creating advertising relevance at scale.

However, it’s worth setting the scene for what’s upcoming by briefly running through the main areas in which AI is currently being used in email marketing:

• Subject lines. With their tireless data-analysis and constant learning processes, AI can generate engaging subject lines which convert at a far higher rate than those produced by humans. But for me, the best part is that you're able to use AI to test in the double figures, a vast variety of subject lines automatically on one send (depending on your mailing list size). The machine will then learn from the interactions and behavior of your recipients and adapt accordingly for the next send. This is something that as marketers, of course, we can, but struggle to do so at scale quickly. This, for me, is where the machine becomes our assistant in driving forward the ideas that we have at scale and provides us with more time to spend on strategizing. Which leads me on nicely to stressing the importance of identifying your email strategy first before embarking on utilizing AI. From experience, the brands that have a clearly defined strategy that then invests in AI technology reap the most rewards.

Dominoes is a brand that did precisely this. 7 years ago, the brand admitted to the world that their product wasn't very good. So naturally, the company re-invented itself. However, new competitors were also taking market share. Their challenge was to increase customer engagement while staying on brand & keep hundreds of franchisees happy. For email, this meant that they needed to remain relevant, reinforce its brand and ensure that they maintained the engagement of subscribers as much as possible. Provided by my phriends at Phrasee, are the individual subject lines they tested:

The very top subject line was the winner and created by the AI tool. Subjects 1-6 were AI generated. Subject line 7 was the human subject line. This may look like small numbers in terms of an extra 5% but for Dominoes that thousands of more openers. For the sceptics amongst you, it's not only driving more opens but also revenue. In only two months, Dominos generated the following:

And this isn’t a one-off case, I’ve witnessed the same for clients. Phrasee also have several case studies on their website detailing the same amazing results as well as the other AI providers available.

Okay so that's quite a lot of detail about AI in subject lines, here are some of the other areas AI is being used:

• Automation. Combining machine learning with automated processes enables AIs to send the right email to the right person at the right time without the need for a human to press the button. Easy to forget that AI powers automation as has done so for some time, to the point now we take it for granted.
• Copy. AI written email copy has been around for some time, but still something that many brands are yet to adopt. Maybe this will eventually lead to AI written website copy based on email interactions?
• Analysis. The analysis is where AI really comes into its own. Nothing is better at parsing digital data than an algorithm made of digital data. Analysis and extrapolation are what makes AIs good at doing everything else on that list – but AIs can also present analysis to us in the form of insights, which are a much-needed element of any marketing strategy.
• Social media. AI written social media advertising copy is something new, and only a handful of brands have started to explore with some outstanding results. Wowcher, for example, reduced their CPL (cost per lead) by 31.2% for utilizing AI-powered copy.

What does this mean for marketers and their customers?

Running an eye over the above, it quickly becomes apparent that AI has very rapidly made itself indispensable to the world of email marketing. Those who have embraced automation and AI are reaping the benefits, seeing a significant uptick in opens, clicks, revenue, and revelling in the heightened exposure of AI optimized content. Those who haven't are struggling to compete with their more digitally advanced competitors.

We're in the midst of what's being called an 'AI Revolution'. And the thing about revolutions is that they bring inescapable changes with them. In the space of a few short years, we've hit that revolutionary tipping point. If you're a company of a specific size, you can no longer choose whether or not to adopt AI in your email marketing. You have to have it, or your emails will sink into the dregs of the inbox. But that's not to say it's just an advancement in the email space, it is also a huge mindset change for your organization. The technology needs to be embraced, and of course, it can be challenged, but at times, there is a tendency to shape it. You need to give it free rein to test and learn without the fear it is going to replace you. Remember, as explained at the beginning of this article, the machine needs us as much as we need the machine.

Here are just a few of how AIs are making themselves an email marketing necessity:

• Most marketing platforms are now designed with a degree of AI and automation in mind. While in theory, you could manually input all your content, all your lists, all your data, then send it all out yourself, analyze each open etc. etc., we've kind of gone beyond that point. Successful email campaigns need at least a bit of automation just to keep the workflow smooth, let alone to collect and analyze pertinent data. Any marketing platform worth its salt will be designed with the expectation that you’re going to be incorporating AI and automated elements. And they simply won’t work as effectively without them. Which, honestly, is excellent. That manual stuff is very annoying for a human mind to have to deal with. Leave it to the robots – they're good at it, it doesn't bore them, they're welcome to it.
• Customers expect the kind of hyper-personalization which only AI can provide. We all know that personalization is the name of the game these days. It's impossible to engage and retain customers without delivering relevant, tailored content which speaks to them on a personal level. Thing is, we as human marketers simply can't get as close to our customers as we need to to build up the personal knowledge that this kind of strategy demands. AIs, however, can analyze and gain insights from customer data on an ongoing basis, enabling them to deliver content tailored to their habits, timetable, likes, and preferences.
• We’re dealing with far more data than humans alone can ever hope to process effectively. Remember Filofaxes? Ah, those were the days. We’re way beyond what even the heftiest Filofax could cope with now. Yet SMEs are currently dealing with virtual tonnes of data every day because that's what's needed to stay afloat in the modern marketplace. The internet makes it easy for us to gather and store data, but sorting, cleaning, analyzing, and using it is another matter. AIs have enabled us to work with so much data so quickly that it's now essential for business survival. Anyone trying to swim in the modern datasea without the aid of an AI will quickly drown.
• Predictive AI is making significant inroads. We're not quite in a 'Minority Report' style world yet, but predictive AI is coming on in leaps and bounds. This shouldn’t be much of a surprise. The greatest predictor of future behavior is past behavior and, as we’ve established, AIs are fantastic at picking up behavioral patterns and using them to great effect. As machine learning advances and more patterns are stored, it’s becoming easier and easier for AIs to accurately predict a customer’s engagement trajectory, and even to pre-empt their next steps. For example, AIs can now predict 'Unsub dates' – the ballpark date at which any given subscriber will lose interest in your content. More to the point, AIs with the right inbuilt automation can take the initiative to re-engage accordingly with precisely what the customer is looking for at that moment. Marketing has always been about staying one step ahead. Now that AIs have pretty much nailed that aspect of things, trying to out-predict them is a bit futile. Working with them is not just the only way forward, it will be the only way to survive as marketers in the near future.

Speaking of the near future…

What can we expect in the future?

Well, as always, it’s tricky to predict the future (but I think I can say with confidence that the future belongs to email marketers who are willing to go big with AI. Further AI integration is inevitable, so we may as well embrace it. If we do it right, the potential could be huge. Here are some of the most exciting areas in which I think that AI could really transform the way that we marketers utilize email as a format and as a channel:

• Greater integration with the IoT – I've not talked much about the Internet of Things here, because it's such a massive topic that, honestly, it needs a page of its own! However, it should be mentioned here. The IoT is transforming how we interact with the internet, and AI has a huge role to play in that. I've spoken before about how voice technology will change email marketing, and many of the points about voice tech apply on a broader scale to AI and the IoT. Through intelligent algorithms, we email marketers will be able to reach people on a broader yet less intrusive scale than has ever been possible before. New formats of communication will be open to us, and the kind of value we'll be able to offer our subscribers will change in ways we can't even begin to imagine yet. The intersection of email, the IoT, and AI is a fascinating area which I would keep a very close eye on in the near future.

All in all, the future belongs to marketers who embrace AI. Which is a scary thing for some of us? We humans have a strange tendency to assign personality and motives where they don’t necessarily exist.

AIs don't have motives, and they aren't competing with us. Alexa may 'talk' to us, but she's not a person. She's an algorithm with a recorded 'voice'. Personally, I think that we need to get rid of the scaremongering about 'intelligent robots' and embrace AIs for what they are: amazing tools which can help us to get the absolute best out of our marketing strategies. For this reason, Explainable AI is the development I'm most excited about in this field. Taking some of the mystery out of the algorithms and integrating them more closely with human operations will, I think, make us more able to get the best of both worlds (human and digital).