How machine learning can impact subscription service offerings
Kristina: Why should subscription-based software companies care about machine learning?
Matt Fleckenstein, SVP Products and Marketing, Amplero: Today’s connected customers demand that companies provide a service to them that is personalized, connected, memorable, and one that provides ongoing value (the more I use it, the more value I receive). Traditional marketing tools simply can’t scale to deliver against these connected customer demands, as rules-based marketing automation solutions require a human to set up complex rules trying to account for thousands of different contexts across the customer journey. Humans simply can’t manage and implement, let alone continually optimize, for thousands of customer contexts across the journey, so marketers are increasingly looking to computers and machine learning for help.
Kristina: How do machine learning marketing capabilities play into the priorities of SaaS subscription companies?
Matt: Because of the rise in subscription-based business models, customer onboarding and retention marketing are, for the very first time in history, more business critical than customer acquisition. (It turns out that the whole “it is cheaper to keep a customer than it is to acquire a new one” saying is very real.) Delivering a personalized, connected, value-increasing service requires subscription companies to know their customers more intimately than ever before and to automate and optimize every interaction across the customer journey. Traditional rules-based marketing automation tools force the marketer to define in advance every part of the customer journey in a very linear, step-by-step sequence of predefined events so that the system can trigger a message when that events occur. The problem/challenge with this approach (in addition to the scalability issues mentioned above) is that customers don’t progress through a journey in a linear fashion on a predetermined course. They are unpredictable — researching a new product on the web one moment and chatting with a support agent about an existing product on their mobile device the next moment. Today’s connected customers hop from one “state/stage” on the journey constantly, not at some preset interval (e.g., wait 3 days and then send message x). By using traditional tools to map out and manage the customer journey, companies often pigeonhole their customers and alienate them by trying to control the process – the complete opposite of what retention marketing efforts are trying to achieve.
Kristina: Describe your typical subscription-based SaaS client and their vision for the future of customer engagement.
Matt: Amplero is the world’s first predictive Customer Lifetime Value Management platform. We leverage machine learning and multi-armed bandit experimentation to enable marketers to achieve what’s not humanly possible. Microsoft is a pretty typical engagement for us. Over the last few years, their Microsoft Office group has worked hard to move away from being a perpetual software company to becoming an online subscription service via their Office 365 brand. They are looking to leverage Amplero to help increase retention/reduce churn. They do this by engaging customers across the entire lifecycle, ensuring that they are onboarding customers appropriately, driving breadth and depth of usage, and renewing (or if necessary, winning back) their most loyal customers. Rather than having their marketers configure hundreds or thousands of elaborate rules that predetermine who should get what message, they are leveraging our machine learning marketing capabilities to let computers figure out who to target with which message on which channel at what point in time. It frees up the marketer to focus on what they most love about their jobs (and coincidentally what computers aren’t very good at doing) — designing rich, captivating creative that engages their customers and drives long-term customer lifetime value.
More from Matt and Amplero tomorrow, including how existing subscription services are already using machine learning to compete.