This repo holds the exercises I did for my product management course. They are practice pieces. I publish them so I can show how I think, and so I can look back at what I learned.
My contribution
I did each exercise and wrote it up. These are course deliverables from the Masai x iHUB DivyaSampark IIT Roorkee programme. They are not products I shipped or work from my job.
The five published sessions
Each published session has its own folder. Session 1, PM with AI, is a PM ecosystem infographic. Session 2, Day in the life, is an infographic on a product manager's day at Cars24.
Session 3, Research methods, is a one-page research plan for Blinkit cart abandonment. Session 4, User interviews, is an AI research prompt tried on Xbox Game Pass. Session 11, GitHub Copilot, is a Snakes and Ladders game with a PRD.
Reading the data first
The research session changed how I would start. Secondary research comes before primary research, because existing data tells you who to talk to and what to ask.
It also taught me that what users say and what users do can differ. My Blinkit plan puts secondary data first, then a hypothesis, then a defined group of users. It uses two methods, set up so that a negative result would change the decision.
Letting AI draft the questions
For the interview session I ran a structured prompt on Xbox Game Pass. The AI first played back its read of the customer and the journey, and I corrected it. Only then did it draft research questions.
It flagged some of its own questions as leading, such as one that assumed playing through Game Pass is difficult, and rewrote them. The prompt gave me a competent first draft. The judgement was still mine to apply.
A small build with a PRD
Snakes and Ladders is a one-file browser game for 2 to 4 players, with an optional computer player. It has 10 snakes and 9 ladders, an exact roll is needed to finish, and a 6 gives another turn.
The game comes with a PRD, and I used GitHub Copilot for the build. The repo lists what I tested, including snake and ladder jumps, the overshoot rule and a full four-player game.