Navi: Structuring a Vague Retention Brief into 200-Plus User Calls
Turned a one-line retention brief into a structured research program, then ran experiments that moved the retention metric.
Krishnam Gupta
APM at Ola Electric
From their time as
Product Management Intern
Navi • 2026 - 2026
Overview
At Navi, Krishnam was handed a single-line brief: grow retention. No framework, no defined problem, no starting point. His first instinct was to resist jumping to solutions and instead make the problem concrete.
The Story
At Navi, Krishnam was handed a single-line brief: grow retention. No framework, no defined problem, no starting point. His first instinct was to resist jumping to solutions and instead make the problem concrete.
He started by clarifying the outcome: what does retention actually mean here, and what would a meaningful improvement look like? From there, he mapped the current user journey end-to-end to identify where and when users were dropping off. Rather than relying on data alone, he went directly to users.
He ran approximately 30 user calls per week, compiling over 200 calls across the engagement. The calls were structured around a simple question: what is actually happening in the moments before a user disengages? The volume of calls gave him a statistically meaningful signal rather than a handful of anecdotes.
One finding stood out. Users were frustrated by the experience of not receiving a reward on certain transactions, particularly those flagged by CFT checks. The friction felt arbitrary and punitive, and it was eroding trust in the platform's value proposition.
Krishnam structured this finding into a product experiment. The team designed a construct that allowed users to bypass CFT checks for one flagged transaction, effectively giving users a release valve for the friction point. The experiment was run, measured, and the result was clear: retention improved by approximately 8 basis points.