We are witnessing a seismic shift in digital marketing. With only 4% of US iPhone users said “yes” to still being tracked by apps after Apple’s vocal iOS update, and the slow death of third-party cookies, brands and marketers need to find alternatives to distribute their content. And Artificial Intelligence email marketing turns out to be the answer. Let’s find out why!
“Email marketing is dead” – how many times have you seen this kind of a provocative post on LinkedIn or other social media channels?
Even without the latest technological turbulences, this marketing format is a stronghold that keeps being rock-solid regardless of multiple algorithm changes in Google, Facebook, Twitter and the rest of the giants.
We have all being fed with the same old pieces of advice when it comes to delivering the content to target audiences:
- Produce SEO content – of course, yes, but nobody can guarantee that you will rank in the top 10 in SERPs, and even if you manage to get there, you can’t bet on staying there forever.
- Post organically on social media – with the algorithms designed to cut down the reach of those organic posts to push you towards investing in paid campaigns, this is getting harder and harder.
- Invest in PPC campaigns – have you conducted one lately on Facebook? Some weird things are going on with their reporting and data.
Emails, on the other hand, (done right) are not shoving brand message in potential buyers’ faces like social media ads. On the contrary, email marketing strategy is about serving content after recipients’ explicit consent given to the brand. And, according to HubSpot, emails generate $38 for every dollar spent, which translates to a mind-blowing 3,800% ROI!
96% of iPhone users say no to tracking
The conflict between Apple and Facebook went on the front pages after announcing the App Tracking Transparency feature coming with the iOS 14.5. As you already know, Apple decided to empower users with the option to turn off the app tracking. Facebook explicitly protested against this move, but Tim Cook’s company defended their stance on the matter of users’ privacy.
What is the actual user approach? After rolling out this feature, the stats strongly confirm that users don’t want apps to track them. Quoting after Gizmodo:
„Just 4% of iPhone users in the U.S. have agreed to app tracking after updating their device, according to the latest data from Verizon-owned analytics firm Flurry. Worldwide, that figure jumps to 12%, a healthy increase but one that still doesn’t spell great news for companies like Facebook that sell targeting to advertisers by hoovering up user data.”
It’s obvious, that we got the point, where retargeting ads and extracting data from those campaigns will sooner or later be useless. Every marketer should bear in mind, that the message is clear – people value their privacy over getting personalized ads on social media.
No more cookies
Apple’s big move comes in line with Google’s declaration to get rid of third-party cookies in Chrome browser by 2022. Experts have already noticed a decline in their value.
There is no mystery, that cookies allow marketers to get their hands on data to chase users with retargeting ads. But with handing over options to manage the cookie settings, that’s getting more difficult and less efficient already.
What are the alternatives for PPC campaigns based on that tracking? Switching to first-party data by building a relationship based on consent and trust.
And that’s where email marketing comes in. Most importantly, powering it up with Artificial Intelligence makes it even more compelling.
How AI revolutionizes our reality
Before we go deeper into using AI technology to make the email marketing realm more effective, we need to forget about artificial intelligence as an abstract, futuristic concept. This tech is already around us and, in general, is about making machines smart to save our time and optimize our daily activities.
AI overarches such terms as machine learning and deep learning – technology and algorithms powerful enough to solve problems without being intentionally programmed to do it and to learn and get smarter on their own, as a perpetuated process.
The impact AI is having on economics is phenomenal. According to Gartner, in 2021 alone we can expect to see AI augmentation enabling 6.2 billion hours of productivity - which means saving more than 700K years.
To look for more tangible examples of AI coming into play, we can talk about:
- Introducing music therapy in healthcare – reducing patients’ anxiety and pain with AI-crafted individualized playlists.
- Using data and algorithms to rewire football – with the help of AI systems, football coaches can look for patterns, for instance in a way of predicting how a particular player will take a penalty kick.
- An organization “Over the Bridge” that took the advantage of Google’s AI Magenta to create a new Nirvana song to raise awareness about depression and mental illness.
So, as you can see, new AI-powered projects and solutions keep on coming to change the game. Just like in the field of email marketing.
Artificial Intelligence in Marketing
Over the last few years, we’ve been observing an increasing role of AI in the marketing stratosphere. With the boom of the IT industry, e-commerce, SaaS platforms and many more business sectors that struggle to get their slice of the pie, there’s a demand for automation solutions that can take the weight of people's shoulders.
Managers, business owners, marketers – we all find ourselves in a situation when there’s too much on our table when it comes to tasks. And, as humans, we must deal with some limitations, such as lack of concentration, having a bad day, health problems, or private-life distractions – very common especially in the remote-work COVID-19 era.
But it’s not only about letting machines, apps, and platforms do the mundane job and repetitive tasks for us. The real point of making the most of AI in marketing is about utilizing this self-learning technology to make substantial business decisions, based on hard data, instead of a gut feeling.
When it comes to the data generated by AI, it can help in:
- better understanding what kind of content is the most compelling to the target audience
- improving user experience by delivering personalized messages
- predicting users’ future decision making to identify which prospects are most likely to convert (and therefore – it’s worth reaching them with proper communication)
Quoting after Search Engine People:
“With new algorithms, the accuracy of data is anticipated to get more efficient. Predicting the ROI and determining sales forecasting will inevitably become a lot more convenient in the future through these innovations.”
Why should you go for AI email marketing?
In essence, the reason AI has become an ally for email marketers boils down to the ability to deliver relevant and personalized content to your (properly identified) target audience at the most adequate time.
Instead of pursuing guesswork and trying to send emails on Monday mornings, Friday afternoons, or Sunday evenings, and getting frustrated by testing the best hours (and not seeing absolute winners), you can get algorithms to sort it out for you.
Debating about whether the graphic header A vs the graphic header B will appeal to your subscribers can be a dead-end, because of personal preferences. On the other hand, this decision can be unbiased and data-driven.
The same goes for copywriting – the subject line, body copy, and the CTA – have an enormous influence on key metrics – open rates and CTRs. Also here, the role of AI is to go for the most optimized version of the copy.
And most importantly – email marketing campaigns can be put on autopilot. Using historical data, users’ interests and behaviour, AI enables to fire away automated emails matched to segments of users and the stage they are actually in when it comes to the customer journey.
What’s also worth highlighting – automated emails are far from being spam. Smart Insights has the point here:
“Analysis of huge sets of data can help you avoid bombarding subscribers with mass generic emails in the future. With personalization and unique content as your point of focus, you’ll be able to stand out from all the clutter they receive from other marketers on a daily basis.”
How does it work in practice?
Here at iAge Technologies, we have been implementing AI solutions for several companies. Taking those case studies under our microscope, we can see some patterns. It turns out, that those businesses already have some datasets gathered through different sources and they’re willing to enter them into an AI system to start work on and help to predict future actions and, therefore, make crucial decisions.
Segmentation
Now, without knowing precisely who to address with brand and commercial messaging, marketers are basically talking to themselves. Creating buyer personas is to some extent a way of impersonating who you’re talking to, but AI takes it further.
Thanks to this technology, you can create a single customer profile that unifies all the attributes and sync that data with other sources.
The reason why this machine learning-driven segmentation leaves old-school personas behind is that it keeps on updating by noticing changes in user behaviour in real-time.
In other words, in a fast-changing world, users' preferences and likeability to purchase something from you can matter today or this week, and be lost right after. That’s why AI comes here to capture the momentum.
Scoring
Earlier on, we mentioned the AI solutions that can help football coaches to know how a particular player is most likely to strike a penalty kick, based on his historical attempts. The same logic applies to predicting a potential buyer’s decision. And this is where scoring models come into play.
To create them, you would need to come up with:
- previous examples of new user profiles
- historical data about their activities
- and data about whether or not your new subscribers hit the CTA button
The system scores those subscribers on a scale from 0 to 100. This scoring enables you to come up with highly targeted campaigns for particular segments of users, with a predictable outcome.
Let the numbers do the talking. For one company, after implementing the scoring model and using the same creatives and copy in the campaign, we’ve found that the top 20% of subscribers identified by the algorithm had almost twice higher open and click rates.
Customer Lifetime Value
In terms of digital business, acquiring new customers is 6.7 times more expensive than retaining the existing ones. That’s why maintaining a relationship and making customers stay with you for longer should be your north star.
Loyalizing customers is what the Customer Lifetime Value metric is about. Luckily, AI makes the CLV predictable. Operating on behavioural data, statistical models and machine learning algorithms, the technology pulls back the curtain and lets you see:
- which subscribers are the most valuable
- what channels are the most efficient and worth investing
- what kind of content hits the nerve
Building a CLV model that will help you to optimize your future email marketing campaigns is worth a shot. Coming back to the previous case study, the predictions for that company were 92-95% accurate, and in the end – those data-driven AI campaigns increased their revenue by 8,7%.
Recommendations
Once you have your email subscribers segmented and built a CLV model, it’s time to personalize your messaging to make your prospects feel they’re special.
For the sake of providing a unique user experience, here at iAge we use the Singular Value Decomposition (SVD) algorithm.
It follows a similar logic as Netflix recommending series or movies based on what you have watched already. If you binged watched all the seasons of House of Paper, you would probably see the Sky Rojo as your possible next stop.
SVD analyses the content in terms of how engaging it was for the subscribers. The better the rating, the higher chance it will be used in the upcoming campaign.
Delivering interesting and fresh content that aligns with users’ interests hugely counts for building relationships between them and your brand, and in the long run, customer retention would benefit from it.
Multivariate testing instead of A/B testing
The last aspect that makes AI email marketing such a powerful tool is multivariate testing. In opposition to the traditional A/B testing, where you can see which of the two options performs better, with MVT, you can juggle multiple elements that can influence the KPIs of your campaign.
So, let’s say, your copywriter came up with five subject lines, and you like each of them. A designer sent you three graphic headers, and you’re missing the “A-HA moment”. The same goes for CTAs.
To give you an example, by creating 108 email combinations coming from crossing different variables, and sending 600K emails, we saw the open rates went up by 22%, and CTRs by 26%.
Time to recalibrate
Wrapping up, the recent revolutionary moves triggered by tech giants will be changing the marketing world as we know it. It’s the best time to be ahead of the curve and to talk about implementing AI email marketing.