Scaled a product to 5,000+ users on KPI-driven decisions
AMIT GAUTAM
Product Analyst. I turn data into decisions.
SQLPYTHONA/B TESTINGPOWER BIMIXPANEL
SCROLL
WEEKLY ACTIVE USERS▲ 35% MoM
5,000+
RETENTION D7
42% → 58%
EXPERIMENT
p < 0.05SHIP
-- which cohort refuses to churn?
SELECT channel, d30_retention
FROM cohorts ORDER BY 2 DESC;
→ whatsapp_referred 2.4x
Experience
Where the numbers came from.
MAY 2026 – PRESENT
GTM Engineer Intern
ClickPost · Remote
0.1% → 47% replies
10 campaigns · ~90% of manual effort automated
MAY 2024 – MAR 2026
Founder & CEO
Distance Connect · New Delhi
5,000+ users · ₹2.5 Cr
35% MoM growth · activation 34% → 52%
AUG 2023 – OCT 2023
Marketing Executive (Growth)
Bubbl Social · New Delhi
1,300 users · 1.7 mo
3.2x viral coefficient · activation +18%
Every number is a person doing something.
Analysis is listening at scale.
I listen, then I ship.
The Loop
Question to decision, on repeat.
01
Question
Start with the decision someone needs to make. Not the data we happen to have.
02
Data
Pull it, clean it, distrust it. Instrument what's missing.
03
Analysis
Funnels, cohorts, experiments. The method matches the question.
04
Decision
A recommendation with a confidence level. Written down, argued for.
05
Measure
Did the metric move? The answer raises the next question.
Numbers I can defend
0+
Users scaled as founder, 18 months
0% → 0%
7-day retention after onboarding A/B
0.1% → 0%
Positive reply rate across 10 campaigns
0K+
Sessions reconstructed in one funnel analysis
Every number here traces to the work in the case files below.
Case Files
Four problems, taken end to end.
CS-01
Zero to 5,000 users, decided by data
5,000+ users
SQLMIXPANELA/B TESTING
Read the case →CS-02
Finding the leak in a 100K-session funnel
68% checkout drop
PYTHONPANDASPOWER BI
Read the case →CS-03
Ship or kill: a 30K-user experiment
p < 0.05
PYTHONHYPOTHESIS TESTINGEXCEL
Read the case →CS-04
Instrumenting an outbound engine
0.1% → 47% replies
N8NCLAYSQL
Read the case →Instruments
QUERY & MODEL
SQL, Python, Pandas
"Modeling behavior, not just tables."
VISUALIZE
Power BI, Excel, Mixpanel, GA, Clarity
"Dashboards people actually open."
EXPERIMENT
A/B, cohort, funnel, hypothesis testing
"Decided by stats, not vibes."
AUTOMATE
n8n, Clay, Zapier, PhantomBuster
"If it repeats, it runs itself."
Need decisions, not just dashboards?
I'm looking for product analyst roles where the numbers ship.