Week 6 Day 2 — Who Do You Choose First?
NST Entrepreneurship · Founder Lab
Week 6 — Day 2/2
WEEK 6 · DAY 2
WHO DO YOU
CHOOSE FIRST?
ICP · Persona · Evidence
Yesterday you chose a customer.
Today you defend that choice.
NST Founder Lab · Entrepreneurship Minor
WEEK 6 · DAY 2
WHO DO YOU
CHOOSE FIRST?
ICP · Persona · Evidence
Yesterday you chose a customer.
Today you defend that choice.
NST Founder Lab · Entrepreneurship Minor
NO NOTES. NO CHATGPT.
MARKET → SEGMENT → ICP → PERSONA
MARKET: ________
Everyone who could plausibly experience the problem
SEGMENT: ________
People sharing meaningful behaviour / context
ICP: ________
The segment we deliberately pursue first
PERSONA: ________
The human pattern we learn through evidence
30 SECONDS.
YOU HAVE 3 SEGMENTS. YOU MAY CHOOSE ONLY 1.
LATE-NIGHT HOSTEL FOOD
A
ALL COLLEGE STUDENTS
B
HOSTEL STUDENTS
C
HOSTEL STUDENTS WHO NEED FOOD AFTER MESS HOURS 3+ TIMES/WEEK AND ALREADY SPEND MONEY OR EFFORT SOLVING IT
₹5,000. 7 DAYS. WHO GETS YOUR ATTENTION?
WHY DID YOU CHOOSE THEM?
PAIN
How badly?
FREQUENCY
How often?
WORKAROUND
Doing something already?
ACCESS
Can we reach them?
ABILITY TO ACT / PAY
Can something happen?
EVIDENCE
Do we know this?
PICK THE 3 THAT MATTER MOST HERE.
OUR FIRST CUSTOMER IS ______
BECAUSE WE OBSERVED ______
✕ because the market is huge
✕ because they would probably like it
✕ because ChatGPT suggested it
✓ BECAUSE WE OBSERVED…
If sentence 2 is empty, sentence 1 is still a hypothesis.
PERSONA
OR
FAN FICTION?
YOU DECIDE.
AARAV · 19
Engineering Student
Loves gaming
Uses Instagram
Drinks coffee
Wants to be productive
Gets distracted
Likes technology
GOOD PERSONA?
FAN FICTION.
WHICH FACT HELPS US PREDICT
WHAT AARAV WILL DO?
DEMOGRAPHICS ≠ BEHAVIOUR
A MORE USEFUL HUMAN
HOSTEL STUDENT
CONTEXT
Mess closes before late study session ends.
BEHAVIOUR
Orders outside food 3–4 nights/week.
WORKAROUND
Combines orders with friends to reduce delivery cost.
SACRIFICE
₹150–₹300/order + waiting.
I DIDN'T TELL YOU THEIR FAVOURITE COLOUR.
Do you need it?
THIS LOOKS
MUCH BETTER.
BUT DID WE ACTUALLY OBSERVE IT?
SPECIFIC ≠ TRUE.
A detailed persona can still be fiction.
WHAT ACTUALLY BELONGS IN A PERSONA?
When / where?
What do they actually do?
What causes action?
EVIDENCE PERSONA
What today?
How often? Time / money / effort / risk?
Where did we learn this?
EVERY CLAIM NEEDS A HOME.
WHICH COLUMN SHOULD EMBARRASS A FOUNDER?
🟢 WE KNOW
Evidence exists.
🟠 WE THINK
Hypothesis.
🔴 WE DON'T KNOW
Unknown.
OPEN YOUR STARTUP.
1 — Write your ICP in ONE sentence.
2 — Write FOUR things you believe about them.
3 — Mark each: 🟢 KNOW / 🟠 THINK / 🔴 DON'T KNOW.
4 — Every 🟢 requires evidence.
No polishing. No designing. Think.
Persona surgery
SWAP WITH ANOTHER TEAM.
CIRCLE ONE CLAIM
THEY CANNOT PROVE.
ATTACK THE EVIDENCE.
NOT THE FOUNDER.
Peer attack
WHO HAS TO SAY YES?
CHILDREN'S INTERACTIVE COLOURING PRODUCT
USER
CHILD
BUYER
PARENT / SCHOOL
GATEKEEPER
PARENT / TEACHER / INSTITUTION
WHO SHOULD YOU INTERVIEW FIRST?
Depends on what assumption you're testing.
YOU INTERVIEWED 30 JOB SEEKERS.
“YES. FINDING WORK IS PAINFUL.”
WHO DID YOU NOT TALK TO?
30 INTERVIEWS CAN STILL BE THE WRONG 30.
Break
WHEN WE RETURN: CAN AI BUILD YOUR PERSONA?
CAN CHATGPT CREATE
YOUR PERSONA?
IT CAN CREATE A PERSONA.
IT CANNOT CREATE YOUR EVIDENCE.
CORRECT PIPELINE
HUMANS
Interviews · Observation · Behaviour
EVIDENCE
What actually happened
AI
Cluster · Compare · Summarise · Spot patterns
PERSONA
Useful human pattern
AI → imaginary human → invented “evidence”
AI ANALYSES. HUMANS PROVIDE REALITY.
WEEK 6 SYNTHESIS
WHAT IS STILL AN ASSUMPTION?
DEFEND YOUR CUSTOMER.
ICP HOT SEAT
MARKET → SEGMENT → ICP → PERSONA
USER → BUYER → GATEKEEPER
OUR FIRST CUSTOMER IS ______
BECAUSE WE OBSERVED ______
PRESENT + DEFEND
ICP MURDER BOARD
NOW TRY TO
KILL YOUR OWN ICP.
Murder board
OUR CHOSEN ICP
EVIDENCE WE HAVE
EVIDENCE WE DON'T HAVE
WHAT COULD PROVE US WRONG?
WHO DO WE TALK TO NEXT?
CIRCLE EVERY CLAIM YOU CANNOT PROVE.
BEFORE YOU LEAVE…
Submit your response
Scan the code or open forms.gle/8Db2hG6phHSURYij9