Term 7 · Module 2 of 4

Who should you found with

Startup Dilemmas Founding or Joining

Practise this module in Claude, ChatGPT or Gemini

Set it up free
Mind map

The Co-founder Problem

Whether you found or join, you are choosing who you build with. A joiner simply makes that choice from outside, before signing anything. The question is not "should I found a startup?" but "who should I build with, and how do I tell before it is too late whether it will work?"

The Co-founder Problem

Case: Zepto / KiranaKart and Ansh Nanda

  • Zepto's public story: two young Stanford dropouts (Palicha and Vohra) in quick commerce reach unicorn status in record time.
  • The omitted part: Zepto began as KiranaKart, and in its early months a third person, student Ansh Nanda, was in the room. By his own account he was deep in founding work: developing the product and technology and lining up early retail partnerships.
  • He says he was verbally promised a 20% stake and co-founder title, conditional on staying through a particular date. He never received the equity. Years later, once Zepto was worth billions, came a legal dispute, a police complaint and allegations on both sides; it was settled privately (terms unknown).
  • No claim is made about who was right. The point: nothing required anyone to be a bad person. It looks like a structural default, what happens unless the founding team actively works against it.
  • Consequence: you cannot simply screen for "dishonest founders" or ask whether someone is trustworthy. You must know what structural signals to look for.
  • Guiding question: if you were Ansh Nanda in January 2021, what could you have seen that showed the agreement was fragile?
  • Co-founding relationships do not fail as random freak events; they fail predictably, for the same reasons, and signals of the fragility are visible from outside even before accepting an offer.

What this module covers

  1. Evidence on team formation: solo vs team, and why founders pick who they pick.
  2. Trust vs complementarity, two things founders confuse, at great cost.
  3. What you can observe from outside the company.
  4. One special category: couple co-founders, a double-edged sword (a tailwind that speeds scaling, or a headwind that suddenly slows it).

It is not a checklist. Team formation is hard to predict even with research: the lecture recounts (as hearsay, unverified) that investors passed on Infosys because they doubted the co-founders would stick, though that team was among India's strongest. Hindsight is 20/20; the aim is to see the risk already inside a founding relationship, to the extent possible, while things are uncertain.

  • Seeing risk does not eliminate it; you may still join. But you walk in with heuristics and a fallback, not blind.
  • Analogy: driving to Delhi in dense fog, the driver followed the road separator closely; a simple heuristic let him navigate what looked like blind driving.

Key takeaways

  • Founding or joining, the real choice is who you build with.
  • Zepto/KiranaKart shows a structural failure mode, not villainy: conditional verbal equity, later entrant, no formal agreement.
  • Trustworthiness screening alone does not work; look for observable structural signals.
  • Heuristics reduce blind risk; they do not guarantee outcomes.

Solo vs Team, and Why Founders Pick Who They Pick

Should you even have a co-founder?

Data consistently show team-founded startups outperform solo-founded ones on average, on survival and scale.

Benefit of a team Why
More hands, skills, capital More resources at the table
Resilience The journey is an emotional rollercoaster; a solo founder has no one to consult or correct mistakes
Removes single point of failure If one person is ill, burnt out or dealing with problems at home, work need not stop
Distributed judgment Two or three people are less likely to be wrong about the same thing at the same time; different experience and angles

Caveats:

  • Better on average; the average hides large variance.
  • A bad team is worse than a good solo founder: energy goes into easing friction instead of building.
  • The team premium appears only when the team actually functions.
  • So "get a co-founder" is not good advice by itself; it is "get the right co-founder, or start alone."

Which seat are you in?

Seat Decision
Founding You actively choose; do it deliberately, not by defaulting to the first available, willing person (e.g. a conference chat, "co-founder speed dating" is called a bad idea).
Joining You inherit the team; you cannot pick co-founders, but you can evaluate whether they picked well.

Homophily

Homophily is "birds of a feather flock together": the tendency to bond with similar people (same language, college, hostel, previous-job team, hometown, school). Analogy: strangers speaking Telugu at a conference in Calcutta bond instantly.

  • Why it makes sense: familiarity breeds trust, you have been tested together (hackathon, project, job), and things are efficient. Nobody runs a structured co-founder search at 2 a.m. before a deadline.
  • Why it backfires: it optimizes for comfort, not complementary skill.
    • Overlapping strengths and overlapping blind spots; agreement is fast and feels great until you need a skill neither has (e.g. sales, finance). Like "taking decisions with a twin."
    • Compresses your network: a co-founder from the same circle adds few new contacts.
    • Survivorship bias: we hear of Flipkart and Meesho (similar-background founders) but not the many that blew up.
  • A third need beyond skills: mission alignment, i.e. why you are doing it and what each is staking.
  • Research: teams of friends have higher co-founder turnover than teams of strangers, not lower. Comfort does not protect a partnership under stress; friends avoid blunt conversations (the elephant in the room: not wanting to be the bad person or the first to raise it).
  • "We have known each other forever" should lower confidence unless you have explicitly discussed who does what and who takes the final decision in conflict, before starting.
  • Homophily delivers speed and comfort early; it does not reliably deliver skill breadth or stability under pressure.

Key takeaways

  • Teams beat solo on average (resilience, risk and judgment distribution), but only functioning teams; a bad team is worse than solo.
  • Homophily = comfort and speed, at the cost of skill gaps, blind spots and network overlap.
  • Friend teams show more, not less, co-founder turnover.
  • Have the roles/decision-rights/conflict conversations before starting.

Trust vs Complementarity

Founders confuse these constantly; separating them matters more than almost anything else in the module.

Trust Complementarity
Meaning I believe you will act in good faith, not cheat me; I feel safe being vulnerable with you You bring skills, networks and judgment I lack; together we cover more ground
Source Shared history, familiarity Differences in capability
Failure if alone Identical skills and blind spots Strangers under pressure

The confusion: you trust someone deeply (say, 10 years), so you infer complementarity that was never tested ("I trust him completely, so we'll make a great team"). The two claims are independent:

  • High trust with zero complementarity: never betrayed each other, identical skills and blind spots.
  • Complementarity with limited trust: cover each other's gaps but remain functionally strangers under pressure. Not automatically a worse basis for a company, but a different risk.

Trust vs Complementarity

The practical test

Would I have chosen to work with this specific person on this specific problem if I had not already trusted them?

  • If the honest answer is no, that is not disqualifying, but then deliberately look for the skill gaps that trust is covering up.
  • Example: Steve Jobs and Steve Wozniak. Wozniak said he simply liked hanging out with Jobs; they were fun and complementary in skills, yet their missions were somehow not aligned.

Zepto revisited

Palicha and Vohra: school students, childhood friends who left Stanford together, about as high in trust and homophily as possible. This probably helped them move fast and skip early friction, but says nothing about the third person, brought in later for a skill rather than prior trust. Trust between the original two did not extend fairness to someone outside the pair. It is a structural blind spot of high-trust, high-homophily pairs (not malice): the newcomer lacks that trust, so terms must be spelled out.

Key takeaways

  • Trust is about good faith; complementarity is about covering skill/network/judgment gaps. Neither implies the other.
  • Test: would I pick this person for this problem without the existing trust?
  • High-trust founder pairs systematically under-protect later entrants.

How Founders Find Co-founders

Research names two team-formation strategies that founders use without realizing they are choosing:

Interpersonal attraction Resource-seeking
Logic Pick someone you like, trust, get along with; relationship first, business logic second Pick someone for what you lack (capital, technical skills, distribution, domain credibility); capability gap first, relationship second
Example Zepto pattern; most early founding "What will you bring?" conversations, easier with strangers than friends
Strength Moves fast, fewer early fights Genuinely complementary on paper
Risk Homophily trap: skill and blind-spot overlap Never tested working together under pressure
Failure image Echo chamber Perfect pitch deck (technical + commercial + domain) that falls apart in year 2 because they never liked each other enough to survive a bad phase

Neither is always right; they carry different risks.

Diagnostic question for a joiner: "How did the founders meet, and why did they pick each other?"

  • Warm friendship story with no mention of skills: probe the skill question yourself.
  • All skills and credentials with no warmth: also worth noticing.

How Founders Find Co-founders

Counterexample: Razorpay

Harshil and Shashank met as young engineering students at IIT Roorkee and ran a student tech group: same college, same circle, high homophily like Zepto's founders. Yet they split roles cleanly and early: Harshil took strategy, external relationships and outward-facing vision; Shashank took product and technical build.

  • This split came not from similarity but from genuinely different working styles and strengths, and honesty about dividing along them.
  • Research point: personality and working-style diversity, even in a homogeneous pair, improves outcomes.
  • The dangerous homophily is not "same college" but "same college and we think and work identically", leaving no one to cover blind spots.
  • Shared adversity helped: both were rejected by over 100 investors, and the company began as a different idea (a crowdfunding platform) abandoned after listening to early users. Plausibly the rejection forced role clarity: when nothing works, "everyone does a bit of everything" stops being viable.

Key takeaways

  • Two strategies: interpersonal attraction vs resource-seeking, each with a distinct risk.
  • Ask joiners' diagnostic: how did they meet and why choose each other?
  • Homophily is not always the villain; undifferentiated roles and thinking are.
  • Pressure that makes alternatives unaffordable can force role clarity.

Evaluating a Founding Team from Outside

What "personality diversity" actually means

Not vague "opposites attract". The useful form is decision orientation:

Co-founding pair that works
  • Outward-facing founder: partnerships, fundraising, market positioning = the company's FACE / story
  • Inward-facing founder: product, technical architecture, operations = the company's ENGINE

For a joiner the question goes beyond "do founders get along?" to: which founder owns the thing I will report to, and does that ownership look real or assumed?

Same-background pairs can work (Razorpay) only when they consciously build complementary roles, often under pressure that makes alternatives unaffordable. It does not happen automatically.

Examples

  • Amagi: Bhaskar, Srividya and Srini are very different. They met in college and built one successful startup, then another that has listed and is doing well, by taking different roles. Bhaskar is CEO (media, people, vision); Srividya works inside on technical matters; Srini does marketing and moved to the US since that is the market.
  • Unnamed NSRCEL startup (Intel-supported business plan competition winners): at the global challenge in the Valley they did not win because their problem was seen mainly in emerging economies and the judges were largely from developed ones. An investor from Iran saw the problem as universal and offered money if one founder stayed in the US and another relocated, equity to be decided later. Founders who had just returned to India for family reasons could not, and the company split, early enough that several unicorns now exist in that space. Lesson: external pressures can shake co-founding balance.

Signals a joiner can observe

Digital footprints (YouTube, podcasts, interviews) help; most cues must be read indirectly, as they are awkward to ask directly. None is a guarantee.

Signal Healthy Warning
How they talk about each other Specific, generous about the other's role; same origin story Vague or minimizing; one founder centres themselves, erasing the other
Role split Visible, named roles "We both do everything"; vague division of labour means roles not settled
Equity Documented, vetted agreement Verbal understanding. Research on tens of thousands of founders: roughly 70% of teams split equity within the first month, often equally, with no vesting; the setup that exposed Ansh Nanda
Disagreement Open, even brusque sparring in public ("we fight all the time"); comfortable venting "We never fight, we've known each other forever" is more concerning
Early third person Talked about respectfully; credited Ignored or minimized
Fit of roles to background Who-does-what matches actual backgrounds Both describe the same job in different words (mismatch)
  • Infosys example: Mohandas Pai joined the founding team after asking sharp questions at the AGM; founders judged his logic right and brought him in at founder level, not merely executive. How founders treat an early outsider signals team quality.

Joiner's perspective: capability over prediction

You may not care whether the team blows up after a year if that year builds capability (sales, marketing, building agents, dealing with investors) you will need later. But do not bet on a team for long without care. At your age you have little to lose, so prioritize roles and skills that will compound, rather than agonizing over judgment calls; such judgments will be needed later and capability should be built when it is possible.

Key takeaways

  • Useful diversity = outward-facing vs inward-facing decision orientation (face vs engine).
  • Ask who owns your reporting line and whether that ownership is real.
  • Observable cues: origin story, role specificity, equity documentation, handling of disagreement, treatment of the early third person.
  • ~70% of teams split equity in month one, often equally and without vesting.
  • Joining can be justified by capability gained even if the team fails.

Interview with Pradeep: Founding, Joining and Betting

Pradeep is a serial entrepreneur and product leader (Mapunity at NSRCEL IIMB; later Yahoo, CommonFloor, Flipkart, Ola, Grab, Swiggy, and the crypto-wallet startup Obvious).

Interview with Pradeep: Founding, Joining and Betting

Career path and the join-versus-start question

  • Graduated 2001 in the post dot-com bust; his Siemens offer was withdrawn. Joined the Karnataka government's IT department (Karnataka Space Research Application Center) as project scientist on telemedicine and distance education (EDUSAT): unplanned, "life has its own plans."
  • Join vs start is a moot point; it depends on your scenario and desired outcome. Two kinds of people: "Will this work for me?" -> a job is probably right; "How can I make this work?" -> startup, because a startup builds into existence something that did not exist.
  • Mapunity: from open-source routing at eGovernance Foundation to a traffic-information idea (a hypothesis: can information drive better choices?). In the pre-smartphone era, cell-tower phone counts served as a proxy for traffic, giving traffic at nodes (e.g. Silk Board) not roads. Outputs included the Bangalore Transport Information System, a bus ETA project, an airport travel-time LED board, and an SMS short code (555).
  • Hedging the leap: with a first child expected and the startup unfunded, he held a full-time job (mornings) and worked on Mapunity evenings for about 6 months, until the first cheque arrived; then he joined full time on a 50% salary cut.
  • Mapunity equity: three-way split (33/33/34, Ashwin 1% higher); nobody thought about what equity meant; the "social lab" mission mattered more. He was chosen by his co-founders; roles were Ashwin (government-facing leader), Pandian (deep technical builder), Pradeep (project manager/coordination).
  • Exit from Mapunity: dissonance over objectives (Pandian, the technical bedrock, left first), plus family economic pressure after the pay cut; a Yahoo opportunity came from winning hackathons.
  • Yahoo -> CommonFloor -> Flipkart -> Ola -> Grab -> Swiggy: moved roles within Yahoo (front-end engineer, developer evangelist, product manager; IIM executive education by luck of a bond-driven vacancy). Laid off when Yahoo onshored the role; found a job in about 15 days through a spreadsheet of target companies and network referrals. At Flipkart he accepted a role cut (product manager instead of senior) and was promoted in about 7 months; his offer was honored at dollar value after a $1 billion round. At Ola he drove as a driver (commercial licence) to understand supply, and built Ola Suvidha, cutting onboarding from half a day to 5 minutes. Lesson: live the life of the user your product serves; niche but essential problems with less competition can make you part of the fabric other companies run on.
  • Start/join also happens inside large firms: changing teams or roles is a mini start-or-join.

North star and mindset

  • North star: be around interesting people doing interesting things; the journey is the reward, not a destination (VP, unicorn). He calls himself an "irrational, impatient optimist."
  • Exploration pattern: try the unknown ("what is the worst that can happen?"), borrow from others' experience (follow what others do, as when driving in a new country), and remember you are not alone (family, friends, well-wishers).
  • Showing initiative: flew to Singapore on a travel visa before any interview and had an offer within days; also the Titan MD anecdote that you must ask ("I want the job").

Managing risk and failure

  • He plans for the journey; outcomes are not in your control, inputs are. He was cash negative for about 18 years.
  • Network is net worth: people underestimate their social capital and ability to get help; ask. When raising for Obvious, asking for feedback led people to offer money, while saying he was raising money got feedback but no money.
  • Failure can be economic or social (both recoverable); the deeper fear is being left alone. Coping resources: network of mentors, an understanding family.
  • Affordable loss: his wife's standard formula, "give me 18 months of money and then do whatever you want"; she also arranged an undrawn education loan as insurance against his leaving a job.
  • No regrets: he reports failing about 47 times, documented.

Obvious (crypto wallet)

  • Left a VP role at Swiggy at 42 (midlife questioning, wanting to build from scratch); co-founders were two ex-Ola colleagues; fundraising came before resigning, showing maturity versus earlier impulse exits. Moved to Dubai because of Indian crypto regulation. A multi-million-dollar acquisition term sheet collapsed three days before closing when the acquirer was hacked; the company shut down. Mistake: "everyone needs a wallet, but nobody is willing to pay for it."
  • As a founder you have more skin in the game: you raise money, are responsible for the team's paychecks (they took pay cuts), and owe investors fiduciary duty and updates, including on pivots.
  • 12 to 18 months: you get a gut feel on whether it will work; by 18 to 24 months, after corrective actions (pivots, fundraises, new people, new markets), you know. The biggest pitfall is the zone of neither success nor failure; "one more fundraise" signals too many eggs in one basket. Co-founders must share ambition; nobody wants to "bell the cat", and the guilt is "am I putting myself ahead of the startup?"
  • Ask co-founders early: what are you in this for, how long, when do you exit, what does the wildest success look like? (unicorn, industry change, a meaningful business, a crafty solution); this shows how deeply each is committed.
  • Quit decision: recommended books Annie Duke's Thinking in Bets and Quit (relates to escalation of commitment).

Five-level bet framework

His move to Obvious was a calculated bet with layered fallbacks; to fail, all must fail:

  1. 1Domain: Fintech
  2. 2Field: Crypto
  3. 3Product: Wallet
  4. 4Bet on myself
  5. 5Each level has a fallback e.g. join PhonePe/Paytm, other jurisdictions, chain funding

(Outcome: wallet and crypto-in-India failed; the self and fintech fallbacks remained.) He contrasts this with Mapunity's "blind" joining.

Bets vs journey

"Will this work for me?" "How can I make this work?"
Orientation Destination Journey
Book Thinking in Bets (Annie Duke): loss aversion; overweighting what worked without knowing why; "winner's luck streak" Shoe Dog (Phil Knight): Nike came from wanting to run and give people good shoes, not building a billion-dollar firm
Reaction to failure Taunts such as "you wanted to be Ambani" (short-lived) Easier to explain: "I'm trying this"

Society, small towns and serendipity

  • Society has "wised up": failure is seen as learning; some investors told him to call when he starts next. Success is partly timing, not just effort; beware survivorship bias.
  • Small-town (non-metro) founders: three archetypes: those who declare "I belong here" (e.g. Mumbai film aspirant), those who ask for help constantly (Pradeep's category: "be shameless in asking for help"), and "silent sufferers" who slowly come out of their shell and let work talk.
  • Luck and serendipity: prepare two insights and a proposal (two pockets at the back, one in front). Insights show domain knowledge (his example: 83% of two-wheelers sold in Chennai in 2025 were electric; about 50% of Karnataka's power from renewables) and create a surface area for serendipity; the proposal is your ask (funding, co-founder, engineer, sale).

Key takeaways

  • Join vs found depends on your scenario and mindset: "will it work for me?" vs "how can I make it work?"
  • Discuss motivation, exit and wildest-success vision with co-founders before starting.
  • Structure risk: hedge the leap, define affordable loss, stack fallbacks (five-level bets).
  • Founders get a gut signal at 12 to 18 months; decide by 18 to 24 months.
  • Network, asking for help and two-insights-and-a-proposal increase luck.

Module Synthesis

  1. 1Choose founding or joining
  2. 2Team or solo?Team
  3. 3How were co-founders chosen? Homophily / interpersonal vs resource-seeking
  4. 4Separate trust from complementarity
  5. 5Check roles: face vs engine
  6. 6Check structure: equity documented, vesting, early third person treated well
  7. 7Decide: join for capability or commit

Key takeaways

  • Founding teams fail for predictable, structural reasons, visible before you commit.
  • Evaluate roles, equity documentation, disagreement style and treatment of early members.
  • Capability gained can justify joining an uncertain team.
Revise before your examAll 3 exam tips from this module in one list
  1. The Zepto lesson is structural, not moral. Verbal promises, conditional equity and a high-trust original pair leave a later entrant exposed even without bad intent.
  2. Match the strategy to its risk. Interpersonal -> homophily trap; resource-seeking -> untested ability to work together.
  3. Core contrasts to retain: trust vs complementarity; interpersonal attraction vs resource-seeking; Zepto (trust without structure) vs Razorpay (similar background, differentiated roles plus adversity); friends vs strangers (higher turnover for friends).

Test yourself on this module

Questions from these notes, one at a time, with feedback. Your weak topics are saved.

Set it up free Takes about a minute. Free plans work.