Case Study: ClassPass lifts email ticket tagging accuracy to 88% with MaestroQA
Key results
The challenge
ClassPass, an online fitness marketplace, launched its first-ever QA program in 2019 but found its grading rubrics too generic to yield actionable data. Only 58% of email tickets were being accurately tagged by agents in Zendesk, which the team called extremely low. Because ClassPass used tagging to understand inquiry volume and measure voice of the customer, the unreliable data weakened decisions about staffing and chatbot automation.
The solution
ClassPass doubled down on QA best practices in MaestroQA, hired one full-time grader, and adopted a binary grading rubric that evaluated agents on spelling, grammar, brand voice, and personalization for a more granular view of performance.
“Only 58% of our email tickets were being accurately tagged by agents, which is extremely low.”
SMSydney McDowellCX Enablement Lead, ClassPass
The results, in context
Email tagging accuracy increased to 88% — a 30% improvement — and chat ticket accuracy reached 87%, up by nearly 20%. Automating cancellation chats saved the equivalent of 6,250 days of chat time annually, and proactive COVID-19 outreach carried a 96% CSAT versus 87% for other inquiry types, which the team credited with helping maintain 83% customer retention against an expected 61%.