Module-3 Evidence, Interpretation & Inference

Evidence, Interpretation and Iteration

Customer interviews generate raw data — what people say, what they do, and what they don't do. The goal is to interpret that data honestly, iterate based on evidence, and avoid the cognitive traps that lead startups to build the wrong thing.

Quantitative Evidence

When you have a website, a landing page, or any prototype, track metrics that reveal user behaviour. These give cold, numerical signals.

Numbers are context-dependent. A 15% conversion rate might be high when you are still searching for a customer segment, but low when you have already targeted a specific group (e.g., pediatricians).

Qualitative Evidence

Numbers tell what happened; qualitative cues tell how people feel and whether you are onto something.

• Listening for signals – enthusiasm, urgency, unsolicited suggestions (“This is a problem, but not this — it’s that”).
• Do not prompt for expected answers. Let the signals emerge naturally.
• Politeness as noise – especially in cultures with high power distance (e.g., India), people say nice things to avoid disappointing you. False praise is not a signal.
• Commitment, not compliments – look for actions: giving an email, agreeing to a follow-up, paying even a small amount.

Exam tip: Watch what people do, not what they say. An email id or a ₹100 payment is worth more than a hundred “brilliant idea” comments.

Cognitive Biases to Watch For

Every entrepreneur carries biases that distort interpretation. Recognising them is critical.

• Confirmation bias – hearing only evidence that supports your hypothesis. Actively seek disconfirming evidence (prove yourself wrong).
• False positive bias – interpreting polite agreement as genuine interest. In some high–power-distance contexts, people may avoid direct disagreement; follow-up behaviour is more reliable than polite approval.
• Sunk cost fallacy – continuing because you have already invested time, effort, or money, not because the evidence says it is working. Cultural fear of failure makes this worse in India. If the evidence is clear, cut your losses and start fresh.

Testing Hypotheses: Three Outcomes

When you test a specific hypothesis (e.g., “Pediatricians will pay ₹500/month for this tool”), the evidence leads to one of three conclusions:

1. Hypothesis supported – great; test the next assumption.
2. Partial evidence – hypothesis was vague or incomplete; refine it and test again.
3. Hypothesis not supported – reject it and decide what to do next.

Pivot, Persevere, or Abandon

Based on the evidence, choose a path:

A pivot is a change in one or more of the core elements:

• Customer segment – you thought A was your customer, but B shows more interest. Example: Chumbak initially sold magnets to international travellers; after airports charged high rent, they pivoted to Indian youth on Facebook.
• Feature set – zoom in on one key feature or zoom out to a broader problem.
• Business model – who pays may differ from who uses. Example: Big Basket discovered that NRIs (paying customers) ordered groceries for their elderly parents in India (users). They started advertising in the US.

Exam tip: Pivot is not failure. It is a course correction toward a better opportunity. The worst choice is to ignore evidence because of sunk costs.

The Learning Card

Disciplined documentation turns interviews into actionable insights. After each interview, create a learning card with:

1. What we tested – clear hypothesis statement.
2. What we observed – actual results (quantitative & qualitative).
3. Key insights – what changed your understanding.
4. Next steps – what to do next based on evidence.

This feeds directly into the Build–Measure–Learn loop: build a minimal prototype (MVP), measure customer response, learn, and iterate.

Guiding Principles

• Get out of the building. Talk to real users about their real problems – not your friends, family, or mentors.
• Watch actions, not words. A quick exit from an interview is a signal. A willingness to give contact details is a signal.
• Look for commitments, not compliments. “You are the next Steve Jobs” is worthless; “Here is my email” is gold.
• Know your biases. Use tools like ChatGPT to check the data for signs of your own confirmation bias.
• Pivot or persevere based on evidence – not on hope, effort, or time invested.

Key Takeaways

• Quantitative metrics (conversion, click-through, sign-up) give cold evidence; qualitative cues reveal true interest.
• Three cognitive biases sabotage interpretation: confirmation, false positive, sunk cost.
• Hypothesis testing leads to accept, refine, or reject – then pivot, persevere, or abandon.
• Learning cards formalise insights and feed the Build–Measure–Learn loop.
• Trust paying customers above all; be alert to politeness as noise.
• “The quick and the dead” in startups: be clever and alert, not stubborn.

Entrepreneurial Market Validation & Customer Discovery

The Corridor Principle states that opportunities multiply once you commit to a specific path. You cannot see the adjacent corridors until you start moving down the first one. This directly connects to hypothesis testing in venture creation: each experiment opens new testable questions.

The Distinction: Problem-First vs. Solution-First

Most first-time entrepreneurs fixate on a solution (a technology, an app). Experienced founders start by searching for a real, painful problem that a paying customer segment will fund.

The example compares education, family communication, and customer support, and selects customer support because paying customers existed there.

Market Dimensions: The "Paying Customer" Filter

A market is not simply many users. A viable market for a new tech product requires:

• Willingness to pay — users who will exchange money for the solution
• Ability to reach — repeatable channels to acquire customers predictably
• Repeatability — the buying pattern must be replicable across many customers

The guest's key lesson: "India has never been a great market for software products as a buyer." For a software startup, developed markets (US, Europe, Australia) offer higher willingness to pay, established buying systems, and productivity as a top-of-mind concern. This is an unfair advantage for a founder who can build from India but sell abroad.

Customer Discovery: Structured Experimentation

The guest conducted 75–100 calls before starting his second venture. The process:

1. Pick a geography (decide which market you will serve)
2. Identify 2–3 problem statements in that market
3. Call potential buyers through network, LinkedIn, referrals
4. Ask only about the problem: Is this a top-3 problem in your company? Does your CEO/board talk about it?
5. Ask about willingness to pay: $10/month or$1000/month?
6. Build a prototype based on patterns in the answers
7. Test the prototype with the same people: "Would you buy this?"

Use time-boxed experiments with explicit outcomes and reviews — not random tinkering.

The Hill to Die On

Once you have evidence that (a) the market is large enough, (b) customers will pay, and (c) you have an unfair advantage, commit to that specific niche. Resist the temptation to solve all 50 sub-problems — pick 1 or 2 and go deep. The "hill to die on" means ignoring competitor funding news, new entrants, and self-doubt, staying focused on that customer segment.

• Narrow niche example: Bulk e-signature connected to workflows (not generic e-signature competing with DocuSign). Solve one workflow perfectly.
• Expansion happens naturally: one solved problem → trust → customer asks for the next related problem → corridors open.

The Corridor Principle in Action

"Unless you start the first corridor, you would not know what are the subsequent corridors."

Once inside a market with the right dynamics (willingness to pay, buying systems), solving one problem reveals many more:

• E-signature → contract management → obligation tracking → income share agreements connected to bank systems
• Each new corridor is a larger revenue opportunity

The guest observed this at Freshworks: each fundraise doubled or quadrupled their addressable market because they kept discovering deeper needs.

Common First-Time Founder Mistakes

• Over-reliance on network for a new venture (the network may not match the new problem)
• Building a "mother platform" that solves everything day one
• Premature focus on branding/IP instead of getting 10 customers who would stop what they're doing if you left
• Treating a global website as a global company — better to target 2 cities explicitly
• Ignoring the "time to pay" dimension — some markets have usage but no revenue repeatability

Key Takeaways

• Validate paying customer existence before building anything. A market of users ≠ a market of buyers.
• Customer discovery must be structured: talk to 50–100 potential buyers, focus on the problem, not your solution.
• Time-box experiments and review outcomes — don't iterate just because you have energy.
• Pick a hill to die on: a narrow problem in a large market where you have an unfair advantage.
• The Corridor Principle means commitment reveals more opportunities — but only if the market fundamentally works (willingness to pay, repeatable acquisition).
• For tech/SaaS, developed markets (US, Europe) still offer 1000× the buyer readiness compared to India.
• Talent and culture are non-negotiable for survival through crises (e.g., COVID). Early hires matter enormously.