Brad Bong

Escalations were fragmented, with no governance on prioritization and no real org-wide visibility into what was stuck. Cases lived across Slack, Notion, and Google Docs without a single system deciding what to do first — which pushed median resolution to nearly 70 days.

An escalation kicked off when an agent ran a Slack workflow, which posted into the escalations channel and auto-created a Linear ticket carrying severity, revenue tier, and key dispositions. Every morning, Claude CoWork scraped that intake, ran triage, and produced a sprint board ranked by severity, revenue tier, and age — so I could clear fast wins, push support-owned issues back to Support or Engineering, and own comms on the true retention cases.

Weekly median escalation resolution dropped from 69 days in January to 4 days by late March 2026. Leadership gained clear visibility into the pipeline — and learned that only about a third of reviewed escalations belonged in Retention at all.

Churn outreach leaned too heavily on volume. Plenty of cancellation surveys were coming in, but no strong system for deciding who was actually worth calling first — which segments had the highest save potential, or what message gave us the best shot at a save.

A smarter save motion: segmentation by cancel reason, appointment volume, processing volume, and time in product; tighter targeting; then a script and cadence redesign built around higher-quality conversations. I measured it against the metrics that mattered — saves, dials, conversations, and wrap time.

Monthly saves 4×'d (3 to 12) in a single month while outbound dials dropped 51%. Live-conversation save rate improved from 6% to 55%, and wrap-up time dropped by about 60% — proving that better targeting and workflows beat brute-force calling.

The Director of Education wanted to grow GCC and APAC but had no clean view of educator CAC or what real expansion would require. We were training educators and talking about expansion, but there was no simple answer to “if we put more educators on the ground here, what does that look like in revenue?”

I built the unit-economics model for those regions. From raw training spend I put educator CAC at $1.4K. Then I combined business development's market heat maps with the economics of a course — modules around $350–500, ~18–30% of course income back to Balanced Body, and three volume bands per educator (6, 24, and 40+ modules a year). Running conservative, base, and aggressive ROI models gave me revenue per educator in each band and a clear recommendation: 13 new educators in GCC and 13 in APAC.

For GCC and APAC together, the base case modeled to roughly $440K/year in education-licensing revenue from 26 new educators, with conservative and aggressive scenarios bracketing that number. Leadership walked away with a real CAC figure, clear headcount targets, and revenue ranges they could plan against — not a top-down guess.

Stripe only told us who had already left. We were closing churn tickets, but there was no simple way to see “this logo is about to walk” until the billing cycle caught up.

An intent-based churn signal off live partner conversations. Any time a partner said or signaled they were leaving, I tagged it, tracked it, and tied it back to account value — so we had a running view of logos that were functionally gone before Stripe said so.

Leadership got an earlier, cleaner signal on churn instead of waiting for Stripe. They could see which logos were at risk, how big the hit was, and where to focus save work before it showed up in the official churn report.