Want to keep your high-risk customers from heading out the door? Well, Framed Data wants to give you a hand.
The company, which today announced that it has raised $2 million from Google Ventures, Innovation Works, Jotter, and NYU Innovation Fund as well as a number of angels, applies machine learning to ending subscription churn. It’s focused especially on B2B SaaS companies, uncovering less than obvious behavioral traits of churned users and applying that knowledge for future use, says director of marketing Tim Wu.
“Churn is the biggest pain we can tackle,” he says, providing the company a way to distinguish itself in an increasingly commoditized data analytics space. And in the B2B SaaS market, “the lifetime value of customers is really clear – they know that if customers leave, they lose x dollars.”
Framed Data also wants the platform to be extensible to predict other outcomes, such as discovering patterns in data that can alert customers that they’ve got a buyer on the line. “For instance,” he says, “that way you can help predict that someone using a company’s free service wants to upgrade to a premium fee-for-service plan.” Framed Data can leverage data such as users’ demographics, the actions previous users took before upgrading, and how they worked with the online service just before upgrading to help it determine from a thousand users the 100 most likely to upgrade. “It’s just a higher-quality way to qualify and warm up leads,” says Wu.
Currently, customers coming onto the Framed Data service get a simple dashboard of insights its machine learning models determine from their data, which is regularly updated as their own users opt for or out of their offerings. “Say, for example, that someone ends up churning out of your service and did something no other churn user did before,” Wu explains. “That’s all really important information and it gets rolled into the machine learning models over time.” So, as someone uses the Framed Data service over time, the models get more accurate.
The service plots the dates and times of expected user churn – the high-risk users over the next seven days to immediately reach out to, or as far out as 30 days if they want to do something more strategic and contextual on their end. Potentially Framed Data could help automate reach-out for its own clients to at-risk customers. According to a blog posted today on the company’s web site by Framed Data founder and CEO Thomson Nguyen, “the real value is in converting these insights [when users are about to leave a product, and why they’re going to do it] into actions that automatically save users from churning.”
Nguyen notes in the blog that the new funding will help the company move further ahead in building the prescriptive arm of its business, to “automatically convert these predictive insights into prescriptive actions.”