Improving Retention Through a Real-Time Data Feedback Loop

The business needed to move from reactive issue resolution to proactive value delivery.

Sling Mobility operated a B2B mobility platform where customer churn was being driven by unexpected downtime and delayed issue detection. While the platform collected large volumes of operational data, it was not being translated into timely, actionable insights, resulting in customer frustration and preventable revenue loss.

Owned product strategy for retention-focused initiatives, definition of the data feedback loop and alerting model, and prioritization. Collaborated closely with Engineering on data pipelines, Operations on real-world failure patterns, and Leadership on retention and revenue targets.

Legacy data architecture with limited real-time capabilities. Operational teams already stretched with reactive workloads. Need to demonstrate impact quickly to justify further investment. This required incremental but high-leverage product decisions, rather than a full platform rebuild.

Analysis of churn data and operational logs revealed a clear pattern: Most customer churn was preceded by repeatable warning signals that were visible in the data but not surfaced in time. The issue wasn’t lack of data: it was lack of productised insight.

Focused on early-warning indicators instead of broad dashboards. Closed the loop between data and action by designing alerts to trigger operational workflows. Measured success via retention and revenue, ensuring vanity metrics like system usage did not drive prioritisation.

I traded a comprehensive analytics rebuild for a narrow, fast-to-ship alerting layer – betting that a handful of high-confidence signals reaching operations in near-real time would move retention faster than a broader dashboard overhaul the team didn't have bandwidth to use anyway. That meant leaving known data-quality issues elsewhere in the platform unaddressed, and it meant the first version deliberately under-served edge cases to keep false positives low, even at the cost of missing some real warning signs early on.

Partnered with engineering to enable near–real-time data processing. Worked with operations to define actionable thresholds. Iterated alert logic based on false positives. Gradually expanded coverage once impact was proven.

Achieved a 17% improvement in retention. Drove a 30% increase in revenue through reduced churn and higher account longevity. Shifted the organisation from reactive issue handling to proactive customer value delivery.