Nitro-Net.com – A Global Marketing Group Company
A quick Google search brings up more than 150 million search results and all those on the first page agree: Artificial intelligence is one of, if not the, hottest technologies of 2019. And yet, relatively few companies are making the most of AI.
Surveying 633 business executives in December, PwC found that just 27% are already implementing AI in multiple areas, while 16% have implemented pilot projects. The rest are still in the planning phase or investigating how to use it.
PwC’s fellow consulting firm, McKinsey, reports that retailers that use AI generate profit margins 10 percentage points higher than their AI-averse competitors. Many marketers hear “AI” and think about chatbots or massive tech companies with seemingly unlimited resources: IBM Watson, Amazon Alexa, Netflix’s extraordinarily sophisticated recommendation engine. In other words, AI is this pie in the sky concept that’s revolutionizing business… but not necessarily theirs.
However, there are plenty of ways AI can enhance one marketing channel everyone uses: email. Here are three of them.
More sophisticated segmentation
Segmentation is the separation of your email subscribers into smaller groups based on common characteristics, often demographics such as age, gender and geography. Determining those segments historically involved manually weeding through data. AI can accomplish this in seconds.
AI also enables email marketers to make create more complex segments around behavioral data, both past and (predicted) future. Who’s likely to make a purchase in the next week? What about the disengaged subscriber who’s on the verge of opting out? These segments call for more nurturing than more generic email communications sent out to, say, all the 25- to 40-year-old men in Atlanta who bought shaving cream last month.
Manually, these segments take a lot of time to identify and test. AI can do that far faster, too, even matching them with the most relevant content.
More powerful personalization
Personalized product recommendations are a well-known AI use case. For one, they’re effective; they drive 26% of ecommerce revenue, according to Salesforce. Netflix’s aforementioned recommendation engine? It helped the company recapture $1 billion in revenue in 2017, largely because it personalizes everything right down to the thumbnails you see on your homepage. Below, Netflix demonstrates how different preferred genres dictate whether the thumbnail for Good Will Hunting features romance or an iconic comedic actor.
AI also enables email marketers to recommend in reverse. Rather than matching people with products, marketers can take merchandise and identify those people who will likely be most interested.
Beyond product recommendations, email marketers can use AI to optimize messages with subject line scoring tools and personalized send time. As of 2019, more than half the world’s population uses email. And the Radicati Group projects that we send and receive more than 293 billion messages every day. If you email someone at 6 a.m. and they consistently engage at noon, why not push your send time back a few hours and decrease the odds your email will get buried in an inbox avalanche?
More adequate automation
Artificially intelligent systems are invaluable in marketing automation. Think of triggered messages, which generate more than three-quarters of email ROI. Welcome emails have unusually high open rates and roughly 10% of cart abandonment emails end with a sale, according to Moosend. They’re all automated… and ideally better than “adequate,” but I had an alliterative theme going with the subheads.
Automation also factors into email send times, provided you’re personalizing them and not going the batch-and-blast route, and frequency. You may send daily emails, but how many of your subscribers engage that often? AI could answer that question, and you can adjust your cadence accordingly.
Assisted by AI, testing can also be taken the next level. Obviously no human can A/B test as efficiently, effectively or as quickly as a machine. That goes triple for bandit testing. Named for gamblers who go to casinos and play several slot machines at once to optimize the payout, bandit testing adds more options to a standard A/B test. On a website, bandit tests simultaneously analyze traffic patterns, diverting to the winning option at any given moment. In an email, that applies to things like subject lines and templates.