Impact

Each number gets room to explain the how, not just the headline. Everything here traces back to work done at Loop and Apna Mart.

40% → 32%

Query rate on reimbursement claims · Loop

The aggregate rate looked steady. It wasn't. I broke the volume down by TPA and disease category and found that a small set of segments was carrying most of the total. Acting on those specific segments — not the average — brought the rate down to 32%.

−21%

P90 turnaround time on slow claims · Loop

Targeted the slowest claims specifically rather than the average. Surfaced disease-wise document patterns across top TPAs, so the teams handling claims could act on the P90 cohort directly. The mean barely moved. That was the point — the slowest 10% is what users actually feel.

3 teams

Product & analytics teams shaped · 2024–Present

GyanDhan, Loop, and Apna Mart — each with a different product surface and user base. Health-tech claims, education lending, and retail growth. The method travels; the details don't.

What I don't count

Dashboards shipped, decks delivered, meetings attended. None of those move a metric. If the analysis didn't change what the team did next week, it doesn't belong on this page.