Founder listens to a kirana owner describe a recent delivery problem on an Indian terrace.

Customer Discovery Interviews in India

16 min read

The short version

  • Ask about the customer’s last real behaviour, not what they might do in future.
  • Never pitch your product during discovery; a polite “sounds useful” is not validation.
  • Interview 15-25 people in one narrow segment, then look for repeated pain, existing spend and a reachable buyer.
  • Count commitments, introductions, pilots, payments and data access, not compliments.

The best customer discovery interview questions are simple: “Tell me about the last time this happened.” “How did you solve it?” “What did it cost you?” “Who approved the spend?” “What have you already tried?” Do not ask, “Would you use my product?” That question produces encouragement, not evidence.

Indian founders routinely spend three months building a product because ten people said it was a good idea. Then they discover that nobody will pay, the user cannot approve the purchase, and the problem occurs only twice a year. That is not bad luck. It is weak customer discovery.

The myth: customers will tell you what to build

They will not.

Customers describe their current pain through the lens of what they already know. They may ask for a dashboard when the real issue is poor data collection. They may request a discount when the real issue is distrust. They may praise your idea because they like you, want to end the conversation, or enjoy discussing new products.

Your job is not to collect feature requests. Your job is to find a painful, frequent and expensive problem attached to a buyer who can act.

This matters even more in India. A user may be enthusiastic while the owner controls the budget. A procurement team may like your product while the finance team delays payment for 90 days. A doctor may want software while the hospital administrator blocks deployment. A kirana owner may complain about working capital but refuse a new app because the current informal system costs less than the switching effort.

Customer discovery is not market research theatre. It is a cheap way to avoid spending your runway on the wrong customer, wrong problem or wrong product.

Customer discovery interview questions that actually work

Use questions that force the person into facts, dates and past behaviour. Start with context. Then examine the last incident. Then quantify the damage. Only after that should you test willingness to change.

Questions about the customer’s situation

  • What does a normal week look like for you in this area?
  • Which part of this process takes the most time?
  • Who else is involved?
  • What happens when the process goes wrong?
  • How often does this problem occur?

Questions about the last real incident

  • Tell me about the last time this happened.
  • What triggered it?
  • What did you do first?
  • Which tools, vendors or people did you use?
  • How long did it take to fix?
  • What was the final outcome?

Questions about money and existing alternatives

  • What do you pay for the current solution?
  • What does the problem cost in lost sales, staff time, delays or errors?
  • What have you tried before?
  • Why did you stop using that option?
  • Who owns this budget?
  • What would need to happen before you could approve a purchase?

Questions about switching and urgency

  • Why solve this now rather than six months ago?
  • What happens if you do nothing?
  • What would make you reject a new solution?
  • Who else must agree?
  • Can you introduce me to that person?
  • Would you share the relevant data or let us run a small pilot?

The last two questions matter. A customer who agrees to an introduction, data review or paid pilot has given you stronger evidence than someone who says your concept is “interesting”.

Investor rule: Treat compliments as noise. Treat changed behaviour, access and money as evidence.

What not to ask in an interview

Bad questions create false confidence. Remove them from your script.

  • “Would you use this?” People predict a future version of themselves badly.
  • “Would you pay ₹999 a month?” You have supplied the price and invited a polite answer.
  • “Do you like this feature?” Likes do not reveal frequency, urgency or budget.
  • “Would you recommend it to others?” They have not used it yet.
  • “Do you face this problem?” Most people will agree with a broad statement.
  • “What features should we build?” Customers are not responsible for your product strategy.

Replace hypothetical questions with historical ones.

Weak question Better question What you learn
Would you use an expense app? How did you submit your last expense claim? Current workflow and friction
Would you pay for faster delivery? Tell me about the last delayed delivery and what you did. Actual urgency and alternatives
Do you need better inventory software? How did you discover your last stockout? Frequency, cost and existing controls
Would this dashboard help? Which decision did you make with the last report? Whether the data changes behaviour

Pick a narrow Indian customer segment first

“Small businesses in India” is not a segment. Neither is “urban millennials” or “healthcare providers”. These labels are too broad to produce useful interviews.

Define your first segment using four filters:

  1. Role: Who experiences the problem every week?
  2. Context: Which city, industry, business size or operating model matters?
  3. Trigger: What event makes the problem urgent?
  4. Buying authority: Who can approve the solution and release money?

For example, “Indian retailers” is weak. “Independent pharmacy owners in Bengaluru with two to five outlets who reconcile stock manually every week” is interviewable.

Do not mix segments in one sample. A founder selling to hospitals should not combine a government hospital administrator, a private hospital CFO and a solo doctor and call the result customer insight. Their budgets, workflows and incentives differ.

At the seed stage, focus beats reach. You need one repeatable pain before you need a large market slide.

How many customer interviews do you need?

Start with 15 interviews in one narrow segment. Continue to 25 if the answers are still scattered. Stop when you can predict the next interview’s story and can explain the differences between customer types.

This is not a statistical sample. It is a pattern-finding exercise.

Fifteen interviews with the exact buyer are more useful than 100 responses from a social media poll. A survey tells you what people claim. An interview lets you inspect the workflow, ask follow-ups and challenge contradictions.

Interview at least three groups when the buying process has multiple participants:

  • the person who suffers the problem;
  • the person who decides whether to buy;
  • the person who implements or supports the solution.

In Indian B2B sales, these are often three different people. Ignore one and your pilot may die after a promising demo.

How to recruit interviewees in India

Your friends, former colleagues and startup community contacts are useful for early context. They are poor proof of demand unless they match the target buyer.

Use warm introductions carefully

Ask for a specific introduction, not “Do you know anyone in logistics?” Say: “Can you introduce me to a fleet owner in Pune running 20-50 vehicles who handles maintenance in-house?” Specific requests get better referrals.

Go where the work happens

For consumer products, speak to people at the point of use. For retail, visit stores. For logistics, spend time at dispatch points. For healthcare, observe registration, billing and records teams. For education, speak to principals, teachers, parents and students separately.

Do not rely only on Zoom. A video call hides the spreadsheet, notebook, WhatsApp group and manual workaround that often reveal the real product opportunity.

Use existing complaint data

Customer support tickets, WhatsApp messages, returns, refund requests, app reviews and sales objections are often better than a clean interview. They capture behaviour under pressure.

Read 50 complaints before writing your interview script. Mark repeated words, delays, workarounds and escalation points. Then ask customers to walk you through a recent example.

Paying for interviews

Do not pay a customer ₹500 merely to say they will speak to you. That attracts people who want the incentive, not people with the problem.

For specialist professionals, compensate their time fairly. A small payment for a 30-minute session is reasonable when it reflects lost working time. But never let an incentive replace evidence of actual behaviour.

Run the interview without pitching

The founder should speak less than the customer. Aim for roughly 20% founder talk and 80% customer talk. If you are explaining your product for ten minutes, you are conducting a demo, not discovery.

Use this sequence:

  1. Set the frame. Say you are researching the process, not selling anything. Ask permission to take notes or record.
  2. Understand the role. Learn what the person owns, how often they perform the task and who else participates.
  3. Walk through the last incident. Keep asking “what happened next?” until you reach the actual workaround.
  4. Quantify the damage. Ask about time, money, delays, errors, lost customers and staff involvement.
  5. Map current alternatives. Identify vendors, spreadsheets, WhatsApp groups, internal staff and informal processes.
  6. Test the next commitment. Ask for data, an introduction, a pilot or a paid trial.

Do not rescue the customer from pauses. Silence feels awkward. It is also where people remember the detail they almost skipped.

When the customer says, “The process is painful,” ask, “What makes it painful?” When they say, “We need automation,” ask, “Which step would you remove first?” When they say, “Our current vendor is bad,” ask, “Why have you stayed with them?”

Contradictions are valuable. If a customer says the issue is urgent but has tolerated it for two years, urgency is probably not real. If they say the current vendor is expensive but renew every year, switching has a hidden cost or the problem is not material.

Analyse interviews without fooling yourself

Founders often collect notes and then select the quotes that support the pitch. That is confirmation bias with a CRM attached.

Use a simple interview sheet. Record one row per customer and separate facts from interpretation.

Field Record
Customer segment Specific role, business size, city and context
Recent incident Date, trigger, steps taken and outcome
Frequency Daily, weekly, monthly, seasonal or rare
Current workaround People, tools, vendors and informal methods
Cost of problem ₹, hours, errors, delays or lost revenue
Existing spend What they already pay to solve it
Buyer Economic decision-maker and approval process
Evidence of urgency Recent purchase, escalation, budget or deadline
Next commitment Introduction, data, pilot, payment or none

Tag each statement as one of three types:

  • Fact: “We lost two days because the invoice was rejected.”
  • Opinion: “A better dashboard would help.”
  • Commitment: “I will share last month’s data and introduce the CFO.”

Facts matter. Commitments matter more. Opinions are hypotheses.

After every five interviews, write down:

  • the three repeated problems;
  • the strongest evidence of payment or switching;
  • the most common workaround;
  • the segment that showed no urgency;
  • the question your script failed to answer.

Do not average unlike customers. If five small retailers complain about stockouts and five enterprise chains complain about integration, you do not have ten votes for one product. You have two possible businesses.

When interviews become validation

Interviews do not validate a business. They produce a decision about what to test next.

Strong evidence has a ladder:

  1. The customer describes a recent, repeated problem.
  2. The customer already spends money or significant staff time on it.
  3. The customer has tried alternatives and can explain their weaknesses.
  4. The customer gives access to data, workflow or decision-makers.
  5. The customer agrees to a defined pilot with a deadline.
  6. The customer pays, even if the first amount is modest.

Do not jump from step one to a ₹50 lakh product build. Build the smallest test that can reach step four or five.

For a B2B product, that may mean manually producing the output for three customers before writing software. For a consumer product, it may mean a landing page, a WhatsApp-assisted service or a paid pre-order. This is why founders should read why building technology before finding customers wastes resources.

In India, a paid pilot can still mislead you if the founder pays from a personal account and the company never adopts the product. Define success before starting:

  • which workflow will change;
  • which metric will improve;
  • who will use the product;
  • who will approve continuation;
  • what the customer will pay after the pilot;
  • when the decision will be made.

Common Indian founder mistakes

Interviewing only English-speaking metro customers

Your first ten interviews may come from Bengaluru, Mumbai and Delhi because that is where your network lives. That can hide language, trust, payment and distribution problems in the next 90% of the market.

If your target includes tier-2 cities, conduct interviews in the language customers actually use at work. Use a local operator when needed. Do not assume a translated survey captures the same meaning as a conversation.

Confusing user enthusiasm with buyer intent

A junior employee may love your product because it removes tedious work. The owner may reject it because implementation takes two days. Interview both.

Accepting “send me the deck” as progress

It is not progress unless the recipient opens a defined next step. Ask: “What would you like to see, and when should we review it?” If they cannot answer, record the lead as unqualified.

Ignoring regulation and trust

In fintech, healthtech, education and employment, the problem may be real but blocked by consent, data storage, licensing, KYC, clinical responsibility or procurement rules. Ask what approvals are required before you build.

A customer saying “we cannot use an external vendor with patient data” is not a product objection. It may be a business-ending constraint.

Using interviews to justify fundraising

Investors do not fund a folder of 40 interview transcripts. They fund evidence that you converted insight into usage, retention, revenue or a credible path to those outcomes.

Your interview notes should improve the product and reduce wasted burn. They should not become theatre for a seed deck. If you are raising, understand the difference between angel capital and institutional capital through this guide to angel investment in India.

A seven-day customer discovery sprint

You can run a useful first sprint this week.

  1. Day 1: Define one segment, one problem and one interview objective. Write down your assumptions before speaking to anyone.
  2. Day 2: Recruit 20 people. Ask for specific introductions. Book at least eight calls or visits.
  3. Day 3: Conduct four interviews. Do not show the product.
  4. Day 4: Conduct four more. Rewrite questions where answers remain vague.
  5. Day 5: Review all notes. Separate facts, opinions and commitments. Identify repeated incidents.
  6. Day 6: Ask the strongest three customers for data, a buyer introduction or a paid pilot.
  7. Day 7: Decide: narrow the segment, change the problem, build a manual test or stop.

Set a kill rule before the sprint. For example: if fewer than five of 15 target customers report a recent problem and none will share data or pay for a test, stop building for that segment. Do not move the goalpost because you like the idea.

Keep a decision log. Write what you believed, what customers showed you and what changed. This prevents the team from returning to the original pitch after a difficult week.

How investors read your discovery work

An investor will ask questions your interview summary often avoids:

  • How many customers had this problem recently?
  • How many already pay to solve it?
  • What did the last five customers do, not say?
  • Who is the buyer and what is the sales cycle?
  • What did you decide not to build?
  • What evidence would prove this segment is wrong?

The strongest founder does not claim certainty. They show a tight learning loop:

Hypothesis → interview → observed behaviour → small test → measured result → product or segment decision.

If your only output is a list of positive quotes, you have not learned enough. If your interviews caused you to kill a feature, change your target customer or abandon a market, that is useful progress. It saves capital.

Investors have seen founders spend ₹30 lakh, ₹1 crore on engineering before discovering that the buyer needed a service, not software. They have seen teams celebrate 100 sign-ups when only three people returned. They have seen pilots run for six months without a purchase decision. Customer discovery cannot fix every business. It can expose these problems before they become expensive.

Bottom line

Customer discovery is not asking whether people like your startup. It is finding out whether a specific buyer has suffered a recent problem, already paid to reduce it and will give you a real next commitment. Ask about the last incident. Follow the money. Demand access. If the conversation produces no data, introduction, pilot or payment, you have learned that the customer is being polite, not that you have found product-market fit.

FAQs

What are the best customer discovery interview questions?

Ask about past behaviour: “Tell me about the last time this happened,” “How did you solve it?”, “What did it cost?”, “What have you tried?”, “Who approved the spend?” and “Can you introduce me to the decision-maker?” Avoid asking whether someone would use or pay for a product they have not seen in practice.

How many customer discovery interviews should a startup conduct?

Start with 15 interviews in one narrow segment. Continue to 25 if the answers remain inconsistent. Stop when you can identify repeated incidents, current alternatives, buyer roles and meaningful differences between customer types.

Should I show my product during a customer interview?

Not during the first problem interview. Showing the product anchors the conversation and encourages feature feedback. First understand the current workflow and pain. Show a prototype later to test whether the proposed solution changes behaviour.

How do I find customer interview participants in India?

Ask for specific introductions, visit customers where the work happens, use industry groups and contact people who have raised complaints or requested help. Recruit the actual user and the budget owner separately when they are different people.

What is the difference between customer discovery and customer validation?

Discovery identifies the problem, customer and current workaround. Validation tests whether customers will take a concrete step such as sharing data, joining a pilot, changing workflow or paying. Positive opinions alone are discovery signals, not validation.

How do I know when to stop customer interviews?

Stop the interview phase when patterns are clear and you can run a real test. If customers describe different problems, use different alternatives or refuse every next commitment, do not keep interviewing forever. Narrow the segment or abandon the hypothesis.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top