How AI in email marketing generated millions in revenue for these brands
Recent breakthroughs in machine learning and artificial intelligence mean more personalized emails and better testing—and early adopters are increasing revenue by millions.
Recent breakthroughs in machine learning and artificial intelligence mean more personalized emails and better testing—and early adopters are increasing revenue by millions.
If it seems like you’ve received an endless stream of email today, that’s probably because you’ve received an endless stream of emails today. 121, to be exact. Granted, you probably didn’t open them all. Studies show that you probably opened only about a third of them.
So, assuming a few were important work emails and a few were important personal emails, the rest of them were from brands. Maybe you clicked a few of those too. And if you did, chances are, it’s because machine learning technology sent just the right message at just the right time.
Email marketers have long been some of the most fastidious testers in the industry, testing subject lines, messaging, time of day and a million other variables to make sure messages are reaching their intended targets. Now, however, machine learning technology can test hundreds of variables at once, and what might have been months of testing can now be done in minutes.
Here are a few brands leveraging artificial intelligence (AI) to make the most of their email marketing.
Mother’s Day is a busy time of year for online flower delivery service Telaflora, and the brand generally sends about a billion email offers a year. To make sense of the mountains of data they were getting from customers’ online behaviors, the brand used Bluecore, a data-driven tool that matches customer data to products, and Custora, a customer analytics tool, to create automated flags that would deliver customized messages for customers at every stage in the buyer’s journey. Teleflora was able to use AI to accurately match the customer to the product they were most likely to be interested in, and the brand saw 50% more orders over the previous year.
Virgin Holidays is one of the most recognizable names for vacation planning in the U.K., selling hundreds of millions in holiday packages each year. Still, when it came to testing subject lines in their emails, the company was having trouble finding a solution that yielded results. So the brand enlisted the help of Phrasee, an AI tool for language generation.
Instead of waiting weeks for copywriters to create subject lines and then for the results of painstaking a/b testing to declare a “winner,” the company fed Phrasee its brand guidelines and voice stipulations. From there, natural language processing (NLP) technology generated on-brand subject lines and then machine learning automated the testing. The results? A 2% bump in sales that amounted to millions of dollars.
Ebay-owned Gumtree is the UK’s number one classified ads site, but increasingly overwhelmed inboxes were affecting the company’s open rates. To spruce up their subject lines, Gumtree also enlisted the help of Phrasee’s NLP and machine learning tool. The result was another win for the AI-generated subject line. Gumtree reported open rates increasing as much as 35%-50% after making the switch.
Virgin Holidays isn’t the only travel company seeing results with AI-driven improvements to their the customer relationship management (CRM). SecretEscapes, a UK-based luxury travel company, has a broad audience—to the tune of 32 million subscribers in 23 countries. The brand makes about 70% of its sales from email marketing, often sending users two emails a day. But SecretEscapes was having trouble finding ecommerce solutions that could keep up with the massive amounts of data generated by churning out 30 deals a day to so many customers from so many different backgrounds.
SecretEscapes enlisted the help of JetLore, an AI prediction platform, to tap into “passive data” around customers’ browsing history, attaching a ranking to each customer based on their interests and then serving them deals most relevant to those interests. The machine learning technology also analyzes data from customers who open emails without clicking, so the brand knows which deals were missed and doesn’t resend. The result was markedly higher open rates and increased revenue.
No matter how much of our testing—or even our writing—becomes automated, email marketing is going to need a human touch for the foreseeable future. However, incorporating machine learning into standard practices (such as testing and analytics) saves time and cuts out the potential for human error, meaning emails are not only more likely to get opened, but more likely to get results.
