B.Tech, Computer Science · UIET, MDU Rohtak
Computer science fundamentals. The base for every technical conversation I have as a PM.
PM at Sportstech · Open to conversations
I build products people come back to.
Product manager with 4+ years across software engineering and product, taking products 0→1 across consumer AI, subscription monetisation, B2B marketplaces and hospitality SaaS. I started as an engineer, and I still check the cohort before I trust the chart.
- Now
- PM, Sportstech
- Before
- Bo3 Lab · DigiStay · Nagarro
- Toolkit
- SQL · GA4 · Mixpanel · Figma
Sportstech Move · D30 retention
+32%
Illustrative shape · 28% → 37% · ≈32% relative lift
AI Coach
3×
daily active sessions per user
Credits, not trials
4%7%
free → paid at 60 days
Black Friday
2.4×
subscription revenue seasonal GTM
- Sportstech
- Consumer AI
- Bo3 Lab
- Subscription monetisation
- DigiStay
- B2B marketplace
- Nagarro
- Hospitality SaaS
- Retention
- Experimentation
- GTM
- 0→1
01 Selected work
Eight products. Every number defined.
Consumer AI, retention and monetisation at Sportstech, plus two 0→1 builds before that. Each case shows the cohort, the window and what I watched so the number couldn't lie.
01 · Sportstech · 2025–26
Consumer AI · 0→1
Daily active sessions per user · indexed
3×
AI Coach
Replaced a booked coaching session with an instant AI coach that became a daily companion.
02 · Sportstech · 2025–26
Retention · AI personalisation
Sample workout week · 30-day user retention +32%
+32%
Sportstech Move
AI-generated goals, streaks and nudges that turned a tracker into a habit.
07 · Bo3 Lab · 2024–25
B2B marketplace · 0→1
Sample top-5 matches · illustrative
+70%
Bo3 Lab marketplace
Built a two-sided Industry 4.0 marketplace from MVP — from first match to PoC. Engagement +70%, onboarding +27%.
08 · DigiStay · 2023–25
Hospitality SaaS · 0→1
Check-in time per guest
<2min
Faster than paper
OCR check-in that non-technical hotel staff learn in 10 minutes; 50+ hotels.
More from Sportstech
Before → after
- 03 Quick Start revampThree entry modes and machine-aware screens for a feature that had gone generic. Adoption · UX 22%32%of MAU · +45%
- 04 Credits, not trialsSwapped a card-upfront trial for monthly credits users spend when they're ready. Monetisation · Pricing 4%7%free → paid
- 05 Churn save flowA reason-routed cancel flow that saves the users who still see value. Subscription retention 72%54%cancellations completed
- 06 Black Friday, feature-ledTurned a hardware-discount event into a software-experience story. GTM · Revenue 2.4×subscription revenue
02 How I think about metrics
Anyone can say retention went up. I'll show you which retention.
"Retention up 32%" means nothing until you know who was counted, over which window, and what else you watched. Here is one cohort counted four ways, and why I use the ones I use.
Fig. 01 — Same cohort, four ways to count it
6/10
counted as retained
Week-4 bracket · days 22–28
Did they work out at least once in week four? Seven chances to land in the window, so I'm measuring habit, not the calendar. This was the headline metric for Sportstech Move.
Day 30, exactly
Who worked out on day 30? Mostly this measures which weekday day 30 landed on. Nobody trains on a fixed calendar day, so I don't use it.
Week-4 bracket · days 22–28
Did they work out at least once in week four? Seven chances to land in the window, so I'm measuring habit, not the calendar. This was the headline metric for Sportstech Move.
Rolling · day 30 or any day after
Effectively a churn check. Same count here, but look at the outlined rows: one bracket user left after week four, one who missed it came back. If bracket moves and rolling doesn't, I've only shuffled people across a boundary. For Move, both moved.
Curve shape · W1–W12
A curve that flattens at 20% is a better business than one at 30% and still falling. Where it flattens tells you if you have a habit. That's why D30 is my operating metric and D90 the honesty check.
Six rules I keep coming back to
Each one shows up in a case study-
01
Find the early action that predicts retention.
At Sportstech it was three workouts in the first two weeks. Once you find it, a retention problem becomes an activation problem, and those are much easier to move.
Seen in · AI Coach -
02
Name the counter-metric before you ship.
It's easy to lift retention by nagging people until they switch off notifications, the channel you'll need next time. Opt-outs were a hard guardrail on Move.
Seen in · Sportstech Move -
03
Pick the window that matches the problem.
Credits were measured at 60 days. Measuring at 14 would have measured the exact problem we were fixing: people weren't ready for premium that early.
Seen in · Credits, not trials -
04
Stated reasons are routing signals, not research.
"Too expensive" from someone who barely opened the app is a value problem, not a price problem. I check what people say against what they do.
Seen in · Churn save flow -
05
No data? Find the smallest version of the answer.
Five conversations, a manual count, a hundred support tickets. Then write the assumption down, so when I'm wrong I know which belief broke.
Seen in · Bo3 Lab -
06
Fix adoption before adding features.
If current users aren't adopting the core feature, more features won't help. At DigiStay the bar was simpler still: faster than paper, or nobody uses it.
Seen in · Faster than paper
03 Journey
Engineer, then MBA, then product.
As an engineer I kept getting pulled toward the consumer experience and the business side of what we built. So I moved to where those decisions get made. Tap a bar to open a stop.
Associate Engineer · Nagarro
Client-facing dashboards and apps for global brands. This is where my technical foundation comes from.
- Lufthansa — front-end UIs (Angular) for a revenue-management dashboard where pricing managers update prices and see revenue impact in real time.
- Cisco ThingQbator — an Electron desktop app that pulls student code from a browser IDE, installs dependencies and runs it locally, so Cisco didn't have to run it on its servers.
- AI chatbot backend — Node.js and Python, connected to the front end over sockets and optimised to a 1.2 s response time.
Executive MBA, Marketing · NMIMS
Done in parallel with full-time engineering at Nagarro. The business half of the picture I kept being drawn to.
Product Consultant (Contractor) · DigiStay
Indian hospitality SaaS. I joined before the company was incorporated and built the MVP with the founders; it scaled to 50+ hotels.
- OCR check-in: under 2 minutes, down from 10–15 on paper.
- Research with 25+ hotel staff and 40+ guests; 8 actionable insights.
Associate Product Manager (Contractor) · Bo3 Lab
A two-sided Industry 4.0 marketplace, built 0→1: onboarding for buyers and suppliers, AI matching, and the path to a proof of concept. I ran the backlog and sprints for 2 developers and 2 designers.
- Onboarding +27%, engagement +70%, retention +10% (2024).
Product Manager · Sportstech
Germany-based connected fitness: smart treadmills, walking pads, strength equipment and the app. I own AI features, subscription monetisation and the experimentation stack, working with design, engineering, data science, marketing and growth.
04 Not on my résumé
Two things a résumé can't tell you.
Urmi MishraProduct Manager · Sportstech, Bangalore
Fig. 04 On site
The unknown excites me more than it overwhelms me.
That's how I've worked across fitness, hospitality and B2B industrial without a background in any of them. I learn a new domain fast, mostly by sitting with the people in it.
- Fitness
- Hospitality
- B2B industrial
I'm easy to work with.
I adapt quickly to new teams and shifting priorities, and I like understanding where different people are coming from: an engineer weighing feasibility, a founder chasing growth, a receptionist with a queue at check-in.
Also — my PRDs spend real space on the problem and the business goal. I want engineers to fall in love with the problem, not just the ticket.
Toolkit
Product
- Roadmapping
- PRDs
- User research
- Experimentation / A/B testing
- Growth
- GTM
- Funnel optimisation
- Pricing & monetisation
AI & analytics
- LLM products
- AI evaluation
- Prompting
- GA4
- Mixpanel
- SQL
- Google Tag Manager
- MS Clarity
- Heatmaps
Tools
- JIRA
- Wrike
- Figma
- FigJam
- Miro
- Postman
- Airtable
- Zapier
Technical
- JavaScript
- React / Next.js
- Node.js
- Python
- Angular
- Electron