पाठशाला Pathshala · ग्राहक Grāhak, The customer · Lesson 02 · Start

Finding the first twenty people to talk to

The interview protocol is useless without people to run it on. How to book twenty qualified conversations in fourteen days with no brand and no budget, using warm asks, WhatsApp groups, LinkedIn and the walk-in.

Pathshala, The Founder Library · 11 October 2026 · 10 min read

The [interview protocol](/library/the-customer-interview-done-properly) is useless without people to run it on, and the step most founders stall at is not the question list but the calendar: twenty qualified conversations in two weeks, booked by someone with no brand, no budget and no excuse. This lesson is the sourcing plan. It treats the twenty as a sales target and works backwards from it.

It covers why the target is twenty, qualified, in two weeks; the arithmetic of asks and replies that gets you there; the four channels that work in India and the way each one is misused; a worked example in Jaipur; and the fourteen-day calendar as a checklist you can tick.

Twenty, qualified, in two weeks: why those three words

Twenty because the pattern in a tightly drawn segment usually becomes visible around the twelfth conversation, and a share of the people who agree to talk, often a quarter or more, turn out to be outside the segment once you ask the qualifying questions. Twenty booked gives you twelve to fifteen that count. Eric Migicovsky’s advice in his Startup School lecture on talking to users is to begin with one or two people and improve the process as you go; the sprint is how you get from two to twenty without it taking a quarter.

Qualified means three things you can check in the first two minutes: the person fits the one-sentence segment; they have met the problem in the last ninety days; and they can describe what they do about it today. A conversation that fails all three is a pleasant half hour and a row to strike from the ledger.

Two weeks because a sprint has an end and a rolling habit of “always talking to customers” never produces a table that can be read. Fourteen days is long enough for referrals to compound and short enough that the founder is still sending asks on day ten rather than waiting for replies.

The arithmetic before the asks

Work backwards from twenty. Every channel has a reply rate and the honest planning figures for a founder with no brand are these: warm asks through an introduction convert about one in two; a direct message to someone in a WhatsApp group you are both in, about one in four; a specific LinkedIn message to a stranger, about one in ten; a cold email, one in twenty on a good week; a cold walk-in at the right hour, about one in three. These are planning assumptions a founder repeats, not statistics, and the point of the ledger is to replace them with your own by day four.

A plan that reaches twenty looks like this. Sixteen warm asks produce eight conversations. Twenty WhatsApp messages produce five. Forty LinkedIn messages produce four. Twelve walk-ins produce three. That is eighty-eight asks for twenty conversations, or about nine asks a working day for ten days: a founder’s morning. Pete Kazanjy’s Founding Sales gives the same order of magnitude for first prospecting, a working list of fifty to a hundred accounts built by hand, and his warning applies here too: build the list yourself rather than buying one, because a bought list is modelled on someone else’s customer.

The number founders get wrong is the first one. They send four warm asks, wait a week for replies, and conclude that nobody wants to talk. Send all sixteen on day one.

Warm asks and WhatsApp groups

Migicovsky’s route in is unglamorous: test the questions on yourself, then friends and colleagues, then the people they introduce you to. In India the warm network is wider than founders use. The chartered accountant whose clients are all in your segment. The distributor you buy from, whose customers are your customers. The college friend in a sales job who visits forty shops a week. The family doctor, the school principal, the society secretary. Each of these is not one conversation but a door to ten, and the ask to them is different: not fifteen minutes of their time but two introductions.

The script is three sentences and fits on a phone screen. Who you are in one clause. The one specific thing you are trying to understand, phrased as their problem and not your product. Fifteen minutes this week, with two time options. No link, no deck, no “we are building”. The moment the message describes a product the recipient has to decide whether they like it, and the polite answer to that is silence.

The second channel is the one Indian founders underuse. Most Indian segments already have a room. The district chemists’ association, the apartment residents, the alumni batch, the distributor’s retailer group, the Shopify sellers of one city, the HR managers of one industrial estate. A WhatsApp group holds up to a thousand people and the segment has usually filled several. Getting added is a warm ask to one member; being useful once you are in is the work.

Two rules. First, read for two days before posting. The person who asks the group the question your product answers is the warmest lead you will find this fortnight, and the right move is a direct message to them that quotes their own question back: “you asked about expiry returns on Tuesday, I am trying to understand exactly that, fifteen minutes?” Second, when you do post to the group, post one specific ask for a conversation, never a survey link. A survey in a WhatsApp group gets the forty people with the least to do; a request for fifteen minutes gets the three with the problem.

LinkedIn without the spam, and the cold ask in person

LinkedIn passed a hundred million members in India in February 2023 and it is the one channel where a founder can filter strangers by title, city and company size before asking. For a B2B segment it is the fastest way to build the forty names. The yield is low because almost every message a buyer receives there is a pitch, so the one that is not a pitch stands out.

The message that works has three parts: a sentence that could only have been written to this person, something from their profile, their company’s hiring page or a post they wrote; the one problem you are studying, in their vocabulary; and the fifteen-minute ask. Connection requests with the note attached do better than cold InMail, and LinkedIn caps how many invitations a free account can send in a week, so the cap is the budget and specificity is how you spend it. Commenting on a few of a person’s posts in the week before you message them makes your name familiar when the message lands, and costs ten minutes.

The fourth channel is the oldest. Paul Graham’s Do Things That Don’t Scale opens with the observation that the most common unscalable thing founders must do is recruit users manually: you cannot wait for them to come to you. Airbnb’s founders went door to door in New York. Migicovsky tells of a YC company selling to firefighters whose cold emails produced nothing and whose founders then walked into fire stations and came out with dozens of ten-to-fifteen-minute meetings.

India rewards the walk-in more than almost any market. The shop, the clinic, the workshop, the CA’s office and the wholesale market all have a quiet hour, and the owner is on the premises. A founder who arrives at a pharmacy at half past eleven, when the morning rush is over, and asks the owner how he handles stock that is about to expire will usually get fifteen minutes and a chair. The same founder emailing the same pharmacist gets nothing. Timing is the whole skill: not month-end for anyone who does accounts, not the first hour for a shop, not Monday morning for an office. Ask the first person you visit when the quiet hour is and the rest of the week books itself.

Nobody is too busy for fifteen minutes about their own problem. They are too busy for thirty minutes about your idea.

A worked example: twenty chemists in Jaipur

A founder wants to build a tool for standalone pharmacies that tracks stock about to expire. Day zero: the segment is owner-run single-outlet pharmacies in Jaipur doing ₹15 lakh to ₹40 lakh a month in sales, managing stock on a register or a desktop billing package. The three qualifying questions: when did you last write off expired stock, how often does it happen, roughly how much was it worth. The ledger is a sheet with sixty rows.

Warm column: the family doctor, the pharmacy the founder’s family buys from, two college friends who are medical representatives and between them visit sixty chemists a month, and an uncle’s CA with eleven pharmacy clients. Fourteen asks on day one; by day three nine have agreed and the two representatives have each promised three introductions. WhatsApp column: one of the representatives adds the founder to a district chemists’ group of about four hundred members; two days of reading surfaces three owners complaining about a distributor’s return policy on near-expiry stock; direct messages to all three produce two conversations. LinkedIn column: distributor sales managers and pharmacy chain category heads in Rajasthan, thirty messages, four replies, three conversations, one of which is a distributor who offers introductions to his best retailers. Cold column: three walk-ins a day between eleven and one for four days, twelve visits, five conversations.

By day ten the ledger has twenty-three conversations done and seventeen qualified. The six that failed were two chain outlets with a central purchase team, three shops too small to carry stock that expires before it sells, and one owner who had moved to a wholesale model. The seventeen say write-offs run between ₹8,000 and ₹25,000 a month, that it happens every month, that eleven of them track expiry in a notebook or not at all, and that the distributor’s return window is the real battleground. The product has moved before a line of code: not an inventory system but a fortnightly list of what to return and to whom. And the segment sentence has changed to exclude chains and the smallest shops, which is the normal and useful result of a first sprint.

The disqualifiers and the ledger

Not everyone who says yes counts. The first two minutes of every conversation are for the three qualifying questions, asked plainly, and if the answers are “never”, “rarely” and “not much” the right move is to thank the person, ask who they know with the problem, and finish early. Politeness in both directions is cheaper than a wasted half hour, and the referral question works on unqualified people too; they often know exactly who has the problem they do not.

The ledger is the asset the sprint leaves behind. One row per name: channel, date asked, replied, booked, done, qualified, and the next introduction promised. By day seven it tells you which channels produce qualified conversations rather than merely pleasant ones, and by day fourteen it is the list the next sprint starts from and the first draft of a customer list if you build. Founders who skip it remember the channel that felt best, which is reliably the one with the warmest conversations and the lowest yield.

The fourteen-day calendar

Day zero: the segment sentence, the three qualifying questions, sixty names in four columns, the three-sentence ask. Days one to three: every warm ask out at once, the first twenty LinkedIn messages, joining the groups and reading them, the first walk-ins booked for the quiet hours. Days four to ten: the conversations, each one closed with “who else should I talk to about this?”, each unanswered ask followed up once after forty-eight hours, and on day seven a reading of the ledger by channel with the second week’s effort moved to whatever is yielding. Days eleven to fourteen: the count closed from referrals and walk-ins rather than fresh cold lists, every conversation scored on [recency, frequency, cost, workaround and authority](/library/how-to-tell-a-problem-from-a-complaint), the segment sentence rewritten if the rows disagree with it, and a decision with a date: another sprint on the revised segment, or build.

Run it once and the second sprint takes half the effort, because the ledger already holds forty names and the referral chain is warm. Run it every time the segment changes. It is the cheapest fortnight a company spends.

The yields above are planning figures a founder repeats, not research; replace them with your own ledger by the end of the first week. The sources are short and the Migicovsky lecture is worth watching before the first ask goes out.