The state of machine learning today is reminiscent of baseball scouting in 2002 when Billy Beane, the general manager of the Oakland A’s, broke conventional scouting practices by using statistical data to analyze and value the players he picked for the team. It was a borderline heretical idea at the time, with fans everywhere calling him crazy and demanding that he resign. However, it turned out that changing the paradigm of scouting led to great success -- so much, that he was acknowledged in a best-selling book and blockbuster movie.
While some baseball organizations were using statistical data back then, professionals were hesitant to trust the models that predicted player performance more than their own instincts. By changing the mold, the A’s won 103 games that year and showed the world that they could outperform teams that had three times their budget. Even then, the new, data-driven approach didn’t catch on quickly. It ultimately took not only a best-selling book, but owners and managers realizing that if they didn’t implement a more data-driven approach, their competition would consistently outperform them.
So how does this relate to email?
This same shift is occurring in email marketing. Taking a data-driven approach that incorporates man and machine enables you to provide the best experience possible for your customers. However, modern-day marketers are looking at machine learning with the same hesitation owners and managers were 14 years ago. Machine learning can predict a consumer’s propensity to engage and interact with content, comparable to Billy Beane’s statistical analysis predicting the success of baseball players. However, many marketers are still manually segmenting, similar to the way scouts would divide their time and travel to watch young baseball players perform. The moneyball moment for marketers falls in the area of algorithmic inboxes.
Your Gmail inbox tracks customer clicks and behavior, leading it to filter unsolicited emails for dating sites and “free” cruises into spam folders. It’s already a struggle for marketers to get promotional emails front-and-center in the primary tab of a customer’s Gmail account. And, as data continues to evolve and Gmail continues to get smarter, it will be even more challenging to get messages in front of your customers. Soon, if your content isn’t enticing or getting clicks, it won’t even make it into the promotions tab.
Right now, marketers can still get away with sending generic emails, but Google is becoming increasingly algorithm-based with what it shows to consumers. If your email is still constrained by a lack of personalization, your engagement rates will become distressed and you’ll be demoted to spam suddenly, without warning.
First, you’ll begin to see your open and click rates decline. Since your emails won’t see any action, Gmail will assume the customer isn’t interested, and an algorithm will push you out of his or her inbox completely. The drop off will be sudden and you’ll see a quick nosedive in promotional email interactions. This creates a vicious cycle, as recipients will no longer be interacting with your messages and you’ll be pushed out of their inboxes.
How can you ensure your messages continue to be seen?
Marketers are hesitant to adapt to an email strategy that integrates a team and technology approach, similar to how baseball scouters were hesitant to be driven by data rather than only their instincts. But brands that build an email foundation focused on machine learning will set themselves up for future success. They’ll be taking advantage of data organization, machine cycles will get cheaper, and topics will become better tagged. These folks who are taking advantage of algorithmic learning will become more contextually relevant and aware while driving engagement and customer experience. The more you use machine learning, the smarter it gets, and the more value you’ll receive from it.
To learn more about taking a team and technology approach to your email campaigns and creating relevant content for your customers, reach out to Rise.