I find patterns in places they shouldn't exist. Business Analyst at Snabbit. Previously Swiggy. Chasing paranormal distributions by night.
On segmentation, the one-order wonder, and what your WhatsApp group already taught you about user behaviour.
Read →The correct SQL query matters less than how you approached it. The right guesstimate number matters less than the structure you used to get there.
Your conversion rate is 3.2%. That number tells you almost nothing. Here is what to look at instead.
We survived as a species because we found patterns everywhere. The problem is we never stopped.
The metric dropped 40%. You were right to panic. Except you weren't. On sanity checking numbers before you tell anyone about them.
AUC is for you. Revenue per cohort is for them. Both are true. Only one lands in a meeting.
RFM segments tell you where your users are. A prediction model tells you where they are going. Here is how to build one and how to know if it actually works.
A concept that sounds simple until your dashboard and reality stop agreeing with each other.
On starting a blog, working with data, and finding patterns in places they shouldn't exist.