June 29, 2021

Marketing data engine minimizes customer churn using Stratus Data’s predictive modeling

Clearbit is a venture-backed B2B software company that helps users optimize all customer interactions using data. Founded in 2015, Clearbit now serves over 1,500 businesses, including Asana, Okta, Notion and Segment.

As part of their growth strategy, Clearbit recognized the need for a stronger customer retention program. Clearbit’s Revenue and Customer Success leaders worked with Stratus Data to develop a system that predicts customer churn, allowing Clearbit to prevent $260,000+ of potential churn-related losses.

 

Challenge

15 Customer Success Managers take on 600 enterprise customers

After closing a $15M round of funding, Clearbit planned to capitalize on its six-year record of success with an accelerated growth strategy. A key focus was a stronger customer retention program.

Customer Success Managers faced the challenge of tracking, organizing, and making sense of data from 1,500+ businesses and 400,000+ users to inform a CRM strategy. Vice President of Customer Success Luke Diaz hired Stratus Data to design and implement a refined churn prediction process. With over 600 upcoming customer renewals, Clearbit needed a system to identify the customers who were at risk of leaving.

“It became apparent that churn was negatively impacting our critical successful factors. I knew we needed to unpack and rebuild the way we looked at data, but we didn’t have data science resources in-house. Stratus Data had a proven record of success in developing machine-intelligence models. They just seemed to get us right away.” –Luke Diaz, Vice President of Customer Success

“It became apparent that churn was negatively impacting our critical successful factors. I knew we needed to unpack and rebuild the way we looked at data, but I wasn’t sure where to start. Stratus Data had a proven record of success in developing machine-intelligence models. They just seemed to get us right away.”
– Luke Diaz, Vice President of Customer Success

Solution

Stratus Data’s algorithm identifies customers who might leave, up to 12 months before their renewal date

Right away, Stratus Data formulated a comprehensive approach to Clearbit’s machine-intelligence model. An important question was timing: “We challenged Clearbit’s Customer Success leaders to think about when they want their team to take action on customer information,” says Charles Pensig, Stratus Data Founding Partner and Data Scientist. “To ask themselves: ‘How far into the future do I need to forecast?’ This would help us decide how to optimize the system.”

Over six weeks, Stratus Data worked closely with Revenue and Customer Success leadership to analyze indicators such as company attributes, communications, and product usage data. Ultimately, they cultivated 400 potential signals to help predict churn months in advance. The product: a model that scans three years of data to correctly predict whether customers, renewing in 3-12 months, are at high or low risk of churning. The Python notebook is run twice per month, then uploaded to Salesforce to inform the team’s weekly flagged-accounts meeting.

Clearbit can now identify 75% of the customers who are thinking of canceling their subscription, long in advance of their renewal dates. “We settled on two to four months in advance,” says Pensig, “which allows ample time to catch unhappy or disengaged customers before they’re completely lost.”

 

Results

VP states algorithm has saved Clearbit $260K+ of potential customer churn

Clearbit’s Customer Success team has always been skilled at predicting churn in big customers. Right away, Stratus Data’s algorithm generated results that complement those human predictions. The algorithm also identifies churn in mid-level customers, whose behaviors are easy for humans to miss. Just in its first quarter of implementation, the algorithm revealed an additional 40% of at-risk customer revenue.

Now, with highly accurate predictive technology aiding their decisions, the Customer Success team can focus on what really matters: their customers.