3 ways machine learning can help publishers in the entertainment category create better experiences and maximize customer value.
The entertainment category covers a lot of ground… basically anything that is, well entertaining: books, movies, music, television, events, celebrities, reviews, parodies… and the list goes on. One thing most entertainment publishers have in common is the problem of loyalty. An outsize portion of your engagement (and therefore your value) can be attributed to a relatively small portion of your overall audience. So naturally, you want more loyal fans (Duh), but you face two primary hurdles:
- Converting Content Tourists: occasional visitors who click around but don’t engage deeply
- Preventing Churn: active users who engage less and less over time until they drop off your radar entirely.
You can invest in creating more and better content, but you’ll be injecting that content into a hugely crowded marketplace. Given that the Entertainment category is essentially a massive amalgamation of niche markets, you can’t count on great content alone to significantly boost your engagement and conversion stats. Be nice to find a stat or trend piece to reference here…
You can crank up your your acquisition efforts… go out and buy a lot of impressions or leads, but that will still only convert a small percentage of users from tourists to fans, and it does nothing to address your churn rate. (And this approach only gets more expensive over time.)
The Importance of Relevance
The answer is, you have to invest in great content and awareness, but the real key to creating and retaining more loyal fans is relevancy. You need to be able to deliver the content that is uniquely able to satisfy their needs of an individual consumer at the moment of its consumption. This is the kind of experience consumers expect online today: content that is personalized to their needs and delivered to them where they are, and if you can’t deliver that, they’ll find it on Facebook or your competition’s website.
Relevancy converts tourists: It promotes discovery based on that individual’s behavior, leading them to more content that they love. It flips a switch in their brains from “cool article,” to “cool site,” which is the first step towards you becoming their “favorite site.”
Relevancy reduces churn: It’s inevitable that people will lose interest in a topic. Trends shift, new media enters the scene, and people move on. But if you can’t predict what a user will like next, they will become more and more likely to churn.
So how do I get me some of that good relevancy?
You already have the data you need; you just need to put it to use. You need to develop some Predictive Intelligence. Know what content is most likely to engage which users in what way right now. The best predictor of what someone will like next is not some combination of demographic data and preferences, it’s their behavior… the thousands of micro-data points involved in their interactions. This is exactly what Facebook, YouTube, Amazon, Netflix and plenty of other have used to create addictive experiences and highly engaged users.
“Easier said than done,” you might be saying. That’s a lot of data to manage, and achieving that kind of rich profile on individual customers (and individual pieces of content) becomes a problem of scale. There are only so many marketers and so many rules, or triggers, or segments you can create.
Enter Machine Learning
That’s where Machine Learning comes in. It’s like injecting a powerful artificial intelligence into your marketing ecosystem, (Actually, it’s not like that, it is that.) A savvy marketer could notice that a certain behavior leads to more shares and create an automated trigger that says something like, “send email “X” to any member of segment “Y” whenever one of them performs action “Z.” With Machine Learning, you start with the outcome: “I want more shares,” and the system automatically creates, evaluates and optimizes those rules and triggers based on a real-time picture of each user’s behavior. Seriously powerful stuff.
What does that look like in practice?
One of the best applications of Machine Learning technology is in marketing outreach — your emails and push notifications. By layering Machine Learning on top of a marketing automation platform, you essentially create a turbo-charged, self-optimizing engagement machine. The system will look at dozens of variables – send-time, sender, subject line, color, layout, content, etc – and crunch the data to determine what notification to send and when to send it, and it can automatically incorporate the content that recipient is most likely to engage with into the email or in-app message. A relevant message delivered across the right channel can easily yield a 3x increase in engagement.
Another application for Machine Learning technology is on-site optimization. This is way beyond simple, programmatic “related content.” If you know who a visitor is, you can use your data on that individual to dynamically adapt the content on your site to surface exactly what he or she is most likely to engage with in that moment. Even if the user is anonymous, you can use their in-session behavior to automatically tune the content they see as they move through your site.
And it’s actually pretty accessible
The technology that companies like Facebook and Amazon have spent millions to develop in-house is now, to a large degree, available to anyone through SaaS services. It’s actually feasible to use advanced algorithms to match users to the right piece of your content on your website, your app, or your email. Right now it’s a competitive advantage, but pretty soon, Machine Learning is going to be table stakes for successful Entertainment marketers and media publishers across all industries.