The arcade ticket as a savings instrument for first-time investors.
Shourya, co-founder of Fello, walked away from BlackRock's institutional Financial tooling to build a phone-shaped on-ramp for the ninety-five percent of India that has never bought a financial asset. The product takes a wager familiar to anyone who has watched a Gen-Z user spend two thousand rupees on Free Fire skins in an evening: route the same impulse into digital gold and an RBI-regulated peer-to-peer fund, then return the money as gameplay tokens rather than withdrawals. Six hundred and fifty thousand sign-ups, eighty-nine percent of the investing base making a first-time financial transaction, an Entrepreneur First cohort, a Y Combinator batch, and a four-and-a-half million dollar round led by a New York seed firm. The argument underneath is anthropological more than fintech: India has a spending capacity for play that dwarfs its visible savings discipline, and the bridge is built from the play side, not the savings side.
In sixty seconds.
Shourya's product hypothesis arrived from the wrong end of the funnel. Fewer than five percent of Indians hold an alternate-asset or equity position, against a global benchmark of twenty to forty percent. Meanwhile the same demographic spent roughly 2.6 billion dollars in 2022 on paid and fantasy mobile games — in-app purchases, skins, fantasy stakes. The capacity to part with money exists; the discipline to compound it does not. Fello inverts the fintech onboarding sequence and pulls users in through the game door, then converts the deposit into a holding in digital gold or a peer-to-peer fund.
The mechanic is structural, not motivational. Every rupee that enters the platform is invested into one of two non-volatile RBI- or partner-regulated assets — digital gold via Augmont, and a peer-to-peer fund via Liquiloans. The principal is never at risk inside the games; users receive separate gaming tokens that act as the play currency, while the underlying investment compounds at six to ten percent annualised behind the scenes. Eighty-nine percent of Fello's investing users have never invested before, and the average user begins with one month of a 25,000-rupee salary spread across a year at about two thousand rupees per month.
The implications run wider than the cap table. Shourya describes a stack where games provide virality and social proof, financial assets provide the stickiness that retains the user once the play loop fades, and the partnership structure removes the regulatory exposure that would otherwise require a custodial licence. Hiring is interns-first because the early team owns the product; the architecture is microservices on a single Flutter codebase; the marketing thesis is that the most expensive line in any consumer-fintech budget is acquisition through Google and Facebook, and games dissolve that cost into word-of-mouth. The conversation also names the next twelve months — new assets, gamification depth, no plans to scale to two hundred staff.
Where to land in the conversation.
Each chapter opens the YouTube video at that timestamp in a new tab.
Six ideas to carry into your own work.
Mental models lifted from the conversation that travel beyond gamified saving. Each one earned its place because Shourya names it specifically and then describes the mechanism inside Fello that implements it.
The wrong-door on-ramp.
Most fintech apps recruit through the right door: rationalist messaging about returns, inflation, compounding. Fello recruits through the wrong door: the game. The thesis is that demographic capacity to spend has decoupled from demographic discipline to save, and the cheap fix is not financial literacy. It is to invert the entry point and let the discipline arrive after the money does.
Principal sequestration as trust theatre.
The mechanic Shourya repeats most: user deposits never enter the games. They are routed to a custodial partner — digital gold or a peer-to-peer fund — and the user receives a token that is the only thing the games can spend. The architecture is a separation-of-concerns pattern borrowed from payment rails. The marketing benefit is that the platform never has to argue about safety; the design carries the argument.
Stickiness from the asset, virality from the game.
Two retention curves operate on different physics. Financial-asset stickiness compounds — once a user starts investing, the inertia of an accruing balance keeps them on the platform. Game virality decays — even a great game burns out a cohort in months. Fello stacks them: the game pulls in ninety-four percent organic referrals, the asset locks them in once they have transacted. Neither alone produces the curve.
The minimum viable scrappy launch.
Fello shipped publicly with two tabs — an ICICI mutual fund and a Tambola game built in-house. The team rejected the stealth-mode pattern because the feedback loop in stealth is too slow. Shourya names the loop discipline: short cycles, no hiding, no halting the live app while iterating. The Tambola game alone produced a ninety-four percent organic referral rate in the first six months, validating both the virality hypothesis and the build-shipped-iterate stance.
Defining acquisition as first transaction.
Most consumer-fintech dashboards count sign-ups; some count app installs. Fello counts the first investment. The redefinition reshapes the funnel because every upstream step now has to carry the user across the transaction threshold, not the registration threshold. Marketing budgets, onboarding flows, and partner-asset curation all change once the unit of progress is the first rupee saved rather than the first download.
Functional ownership over veteran ballast.
Fello's early tech team was made of interns who converted. Shourya names this as a deliberate move against the get-a-grey-hair-on-board default. The wager is that ownership compounds faster than experience when the product is new, the room for error is structured deliberately, and the founders take problems to the team rather than coaching from the side. The phrase he uses: cultural functional ownership. Every team member runs a piece of the product.
Fifteen things to walk away with.
Each one carries the timestamps where the moment lives and a transferable note for work that is not gamified saving. The order roughly tracks the conversation from market thesis through architecture, hiring, regulation, and the next twelve months.
The 2.6 billion observation that defined the product.
Shourya names the headline figure that gave the team its commercial confidence: 2.6 billion dollars spent in 2022 on paid and fantasy mobile games in India, against fewer than five percent of the same population holding any equity or alternate-asset position. The asymmetry is the product. Customers willing to drop a thousand rupees on a Free Fire skin or fantasy entry fee are not, in any obvious sense, capital-constrained. They are intent-constrained. The intention to spend on play has been engineered into a reflex over five years of free-to-play monetisation; the intention to save has been left to school-curriculum cliches.
The product implication is that the funnel does not need a financial-literacy pitch. It needs a friction pivot. Once a user is on the platform for the game, the deposit that funds the play is also the deposit that funds the asset; the user is saving by accident the first time and by habit by the third. Shourya's pitch to investors leant heavily on this point because it inverts the conventional fintech model where literacy precedes onboarding.
The arcade-ticket analogy is the entire pitch in one sentence.
The clearest articulation Shourya gives, used twice in the conversation, is the arcade. You walk in, you buy a ticket, you use the ticket to play. The ticket is currency, but the rupee it cost you is gone the moment the ticket lands in your palm. Fello redesigns the moment. The rupee does not buy a ticket; it buys an investment unit in digital gold or a peer-to-peer fund, and a separate token is minted that does what the ticket did. The game does not know the user has any money inside it. The investment does not know the user is playing.
The architectural consequence is what Shourya means by principal guarantee. The user cannot lose the rupee inside the games because the games never had it. The investment compounds at six to seven percent on the digital-gold side and ten percent on the peer-to-peer side. The token economy is a parallel ledger, separable, replaceable. The legal exposure is reduced because Fello never operates the fund or holds the gold; the partners do. The product wins by what it refuses to be a custodian of.
Eighty-nine percent of the investing base is first-time.
The metric Shourya offers as proof that the thesis is more than rhetorical: out of the platform's investing user base, more than eighty-nine percent had never invested in a financial asset before they came to Fello. Sign-ups are above six hundred and fifty thousand; downloads exceed that. The number is not customer-acquisition vanity. It is conversion of an audience that was previously closed to the asset class entirely. The platform has not stolen users from Zerodha or Groww; it has manufactured them from the gaming audience the existing fintechs were not converting.
The second-order effect is what Shourya quietly cares about: the cohort, once converted, builds the habit. The average user begins at around two thousand rupees a month on a 25,000-rupee salary, and the trust factor compounds the deposit over the trailing twelve months. The metric to watch is not the headline sign-up number but the per-cohort retention curve. If first-time investors become recurring investors, Fello has built a new pipe into the asset class. If they peak at the first deposit and decay, the gaming layer is doing more work than the financial layer.
Stable architecture means fault tolerance per rupee.
Shourya redefines what scalable architecture means in 2023. The old benchmark was Black Friday or Diwali peak load. The current benchmark, for a consumer-fintech with a gaming surface, is fault tolerance per transaction. Every rupee that enters the platform must be routed correctly, invested correctly, and reconciled against the partner system within a tight window. The platform cannot afford a lapse. Recon checks are layered two and three deep. Microservices isolate the failure domains so that a payment regression cannot break onboarding.
The deeper claim is that the architecture is doing work the marketing copy does not have to. The trust the user feels when their first hundred rupees lands cleanly is what makes the second deposit possible. Fault tolerance is the silent feature; it is also the most expensive engineering line item because it requires Recon infrastructure that does not show up in user-facing screens. The unit economics of trust in this product are paid for in engineering hours, not advertising spend.
The Gen-Z audience is the adversary, not the user.
One of the more honest moments in the conversation. Shourya describes the platform's primary cohort — salaried twenty-one to twenty-five-year-olds who are also heavy mobile gamers — as people who actively try to game the platform. They probe for loopholes, exploit referral mechanics, look for arbitrage between gaming token economies and the underlying asset return. The team's posture is not defensive; it is anticipatory. Every iteration cycle is short, every new feature is shipped with a security layer audit, and the team assumes the audience will find the seam before the QA pass does.
The transferable observation is that the audience-as-adversary frame produces a different architecture than the audience-as-customer frame. It builds in adversarial review at every stage. It privileges quick iteration over comprehensive specification. It treats the user behaviour log as a signal of the product's loopholes more than its product-market fit. The Gen-Z cohort, in this read, is a quality-assurance asset disguised as a demographic risk.
Flutter and microservices as a deliberate trade-off.
The technical stack is a single Flutter codebase that compiles to Android, iOS and web, on top of a microservices backend. Shourya picks Flutter because it lets a small team ship to all three surfaces without forking the engineering effort. He picks microservices not for the buzzword but for the failure-domain isolation: a regression in the payments microservice should not break the onboarding flow. The combination lets a tiny team iterate faster than a comparable team with a fork-and-merge mobile codebase plus a monolithic backend.
What Shourya does not say but the architecture implies: the team is forecasting a wide product surface area. Microservices and Flutter both pay rent in the future, not the present. They are slower to set up than a monolith, slightly more expensive to host, and harder to staff for. They earn their cost when the team adds a fourth, fifth or sixth asset class without rewriting the platform. The decision is a bet that Fello's product surface grows wider than it is today — new games, new financial assets, new partner integrations — faster than a monolithic stack could absorb.
The interns-first hiring stance.
Most early-stage technical hiring playbooks reach for veterans. Shourya took the opposite. A large share of Fello's early tech hires were freshers and recent graduates, several of whom joined as interns and converted. The reasoning is structural: a fresher whose first job is the company tends to own the product as their own, whereas a senior hire arrives with a half-formed view of how things should work. The trade-off is that founders carry more of the day-to-day problem-solving load, because there is no one above them in the technical hierarchy to delegate to.
Shourya names this as cultural functional ownership. Every junior hire runs a function end-to-end — the front-end app, the payments layer, the gaming integration — and is forced to make calls under uncertainty without an experienced person above them. The talent develops faster under that pressure than it would on a senior team. The cost is that mistakes happen, but the early architecture leaves room for them. The decision compounds because the fresher who survives the first eighteen months becomes a senior who knows the product better than any external hire could.
The partner-asset strategy avoids custodial exposure.
Both core assets on Fello are sourced through partners. Digital gold runs through Augmont, a partner that holds the physical gold backing the platform's digital units. The peer-to-peer fund runs through Liquiloans, an RBI-regulated NBFC that aggregates the underlying loan book. Fello operates the user-facing layer, the tokenisation, and the transaction routing — it never holds the financial instrument or operates as custodian. The architecture removes a heavy compliance burden that would otherwise require Fello to be regulated as an asset manager or a fund.
The choice is deliberate and Shourya is explicit about why. A two-year-old company should not be building a fund. The principles have to come first — flow correctness, fault tolerance, user trust. Custodial risk is a category of risk that can sink a young fintech if a single transaction misroutes, and the partner structure transfers that exposure to entities that have already built the operational discipline. The bet is that the company can layer in custodial functions later, when the user base is large enough that the operational margin justifies the regulatory cost.
Acquisition redefined as first transaction.
Shourya gives one of the cleanest metric redefinitions in the episode. Most fintech apps count sign-ups; some count downloads. Fello counts the first investment. A user who downloads, signs up, plays a game and never deposits a rupee is not an acquired customer in Fello's books. The reframing has ripple effects across the marketing organisation: every upstream step has to carry the user across the transaction threshold, every onboarding screen is optimised for the deposit flow, every partner asset is curated for first-time investor approachability.
The metric is harsher than a sign-up count and the funnel looks worse in dashboards, but it forecasts revenue more accurately. A platform that earns a percentage commission on transactions and AAUM cares about transacting users, not registered ones. The discipline of using the metric is to refuse the vanity number internally; the discipline of communicating it externally is harder, because investors are trained to read sign-ups as growth. Shourya works the metric through anyway, because the per-cohort retention curve cannot be modelled honestly against any other definition.
Network effects without paid acquisition.
The number Shourya is proudest of is not the six hundred and fifty thousand sign-ups; it is the ninety-four percent organic referral rate in Fello's first six months. The in-house Tambola launch drove the early growth almost entirely through social diffusion. The hypothesis is that financial products alone produce stickiness but no virality — you do not text your friend about your gold balance — while games produce virality but no stickiness. The combination is the only way for a young fintech to compound without burning marketing budget on Google and Facebook.
The deeper observation is about peer behaviour in financial markets. Crypto users want to invest with their friends. Fantasy-sports players form leagues. The act of saving has been culturally framed as solitary, but the act of playing has not. Fello uses the gaming social layer to drag the financial behaviour into a shared context. Once the user invites two friends to play, the trust transfer extends to the asset side — the friend knows the platform is real because the inviter showed them the game first.
Scrappy launch and the short-cycle loop.
Fello launched with two tabs — an ICICI mutual fund and an in-house Tambola game — against the conventional stealth-mode wisdom in which the founders hide the product for nine months and unveil a finished version. Shourya describes the decision as a refusal to slow down: even after the initial launch the team never paused the live app to rebuild. The iteration cycle was short, the user feedback was immediate, and the team kept what worked while replacing what did not.
The trade-off is that the first three months of the product were embarrassing by the founders' own standards. The early app, by Shourya's account, looked completely different four months before the interview — the colour palette, the design system, the onboarding sequence were all replaced. The discipline is not to defend the early version but to learn from it fast enough to replace it before the user base remembers what it used to look like. The cost of an embarrassing early version is small. The cost of a polished version that is not yet right is much larger.
The 2,000-rupee monthly deposit and the upward spike.
Shourya describes the per-user economics in concrete terms. A user with a 25,000-rupee monthly salary — the early salaried twenty-one to twenty-five segment Fello targets — tends to deposit around 2,000 rupees a month, or roughly one month of salary spread across a year. The figure is not aspirational; it is the observed median in the cohort. As the user's trust in the platform compounds, the deposit number trends upward. The platform is engineered for the long arc of that trust curve, not the first deposit.
The economic implication is that revenue from any single user grows with tenure, not breadth. A Fello user who has been on the platform for eighteen months is several times more valuable than a comparable user at three months, not because their salary has grown but because the deposit confidence has. The platform's marketing dollar should chase tenure-extending behaviours — cohort engagement, second-asset adoption, friend referral — not first-deposit volume. The discipline is to model lifetime contribution, not month-one revenue.
Multi-game listings as a content velocity strategy.
Fello does not build all its games in-house. After the initial Tambola, the team partnered with external gaming studios to license seven to eight games onto the platform. The rationale is content velocity: a gaming surface needs constant freshness to retain the cohort, and the marginal game built in-house competes for engineering time against the core financial platform. Licensing displaces that build cost while keeping the appearance of a steadily growing arcade.
The partnership model carries a different unit economic profile. Each licensed game arrives with revenue share, integration cost, and a cap on customisation. Fello accepts the constraints because the games are not the product — the financial layer is. The games are a recruiting surface, a retention surface, and a viral channel; they do not have to be exceptional on craft, only good enough to keep the user playing. The team's resourcing tells the story: a small in-house team focused on the financial infrastructure, with the gaming layer carried by external studios who specialise.
The next twelve months are scale, not pivot.
Asked about the roadmap, Shourya is unambiguous: the model has been validated and the next year is about scale. Deeper penetration into the existing customer segment, more first-time investors converted, more assets added carefully to the existing two. The team intends to stay lean rather than balloon to two hundred members; the conviction is that strong individuals carrying functional ownership produce more output than a wide horizontal of average ones. Hiring will be conservative; the architecture is intentionally built to scale on infrastructure rather than headcount.
The interesting omission is the absence of a custodial pivot in the next twelve months. Fello is not rushing to build its own fund or its own gold storage. The principles — transaction correctness, fault tolerance, asset diversification — have to compound first. Shourya names this as deliberate caution about handling user money. The implication is that a young fintech that grows up too fast risks taking on operational obligations that exceed its operational maturity. The discipline is to scale the existing model before adding new responsibilities.
Marathon discipline as a founder operating system.
The closing personal note is the most candid moment in the conversation. Shourya talks about Sundays as a deliberate non-work day, books like Chip War by Chris Miller and Bob Iger's Ride of a Lifetime as anchors, and the explicit framing of the company as a marathon rather than a sprint. The early days were strenuous; the current discipline is about sustainability. The bet is that a founder who burns out at year three does not build a year-ten company, regardless of how good the year-one product was.
The deeper observation is what biographies are doing in Shourya's reading list. Non-fiction founder narratives — Bob Iger at Disney, Chris Miller on the semiconductor power dynamic — provide the long-horizon framing that VC-cycle conversation cannot. The week-by-week metrics produce one kind of attention; a fifty-year industry arc produces a different one. The founder operating system Shourya is building is one where both attentional modes coexist, with the daily metrics anchored by the weekly recovery and the weekly recovery anchored by the long-arc reading.
Lines worth keeping near your desk.
The jargon, unpacked.
Some of these are specific to the Indian regulatory and gaming context; some are pan-fintech operating vocabulary. Skim, mark the unfamiliar, come back later.
Check what you actually retained.
Try to answer before you click. The point is to notice where the conversation is fuzzy in your memory, then return to the transcript.
Five prompts. Notes save in your browser.
Use these to convert the reading into your own decisions. Nothing is uploaded; storage is local to this device.
Shourya routes a play impulse into a savings outcome by inverting the on-ramp. Where in your own product is the wrong-door entry hiding — the adjacent category your audience already adopts that you could route them through
Fello redefines an acquired customer as one who made the first transaction, not one who signed up. What is the harsher metric in your own funnel that you have been avoiding, and what would change on Tuesday if you adopted it
The partner-asset strategy keeps Fello out of custodial scope deliberately. What capability are you currently building in-house that should be partnered for at least the next twelve months
Shourya hired interns-first against the veteran-default. What would your team look like if your next three hires came from the bottom of the experience curve rather than the top, and what room for error would you have to architect for them
The marathon discipline — Sunday off, biographies, the long-arc reading — is the founder operating system Shourya names. Which of your current weekly habits is borrowed from the sprint version of your career, and which from the marathon
The strongest version of the disagreement.
Four counter-arguments that an honest sceptic would press on this conversation. Each is written to be persuasive, not to win.
Gamified saving is a regulatory accident waiting to be reclassified.
The thinnest part of Fello's argument is that tokens earned from a financial deposit are not themselves a financial instrument. Indian regulators have moved aggressively against fantasy-sports operators on the boundary between skill and chance, and the GST council has reclassified the entire real-money gaming segment at a punitive rate. The honest read is that any product where money goes in and play comes out is exposed to a future reclassification — PPI licensing, GST on token spend, SEBI scrutiny on the assets-as-game-fuel framing. Shourya’s architecture is elegant today; it survives a tail-risk regulatory event only if Fello can de-couple the token economy from the underlying asset on twelve hours' notice. That capability is not visible in the current architecture and the cost of building it retroactively could be substantial.
The eighty-nine-percent first-time-investor metric flatters because the bar is low.
A first-time investor on Fello is anyone who deposited any amount once. The metric does not test whether the user invests a second time, holds beyond the lock-in, or grows the deposit. The honest counter is that the eighty-nine percent number describes the trial conversion, not the durable investor base. The relevant comparator is not Zerodha’s onboarding (which is high-friction by design) but Google Pay’s gold-savings product or Paytm’s mutual-fund tab — both of which have far larger first-time-investor cohorts that decay rapidly. Without a published cohort-retention curve at six and twelve months, the eighty-nine percent is a directional indicator, not proof of category expansion.
The unit economics of a 2,000-rupee monthly deposit do not pay for the engineering.
Fello earns percentage commissions on AAUM and transaction volume from partner assets. At an average deposit of 2,000 rupees a month and yields of six to ten percent annualised, the commissions are thin. A user who deposits 24,000 rupees in a year and stays for two years contributes commissions in the low hundreds of rupees. Against the engineering cost of fault-tolerant transaction infrastructure, microservices, Flutter cross-compilation and a security-audited release cadence, the contribution per user does not justify the build. The thesis requires either a much higher deposit growth curve than the cohort currently shows, a step-change in monetisation (the gaming partnerships, brand revenue) that is not yet visible, or a structural cost reduction that the architecture has not yet captured.
The interns-first hiring strategy looks frugal until the technical debt arrives.
Hiring freshers who own functions end-to-end is cheap in salary and high in motivation. It is also, predictably, the route that produces the most technical debt — under-tested release paths, brittle integrations with partner systems, half-finished migrations. Once the platform crosses a million users and the partner integrations multiply, the debt compounds faster than the team's ability to retire it. The honest counter is that the moment to hire senior engineers is not when the debt becomes intolerable; it is two quarters before. Fello’s deliberate avoidance of veteran ballast may be the right call at six hundred thousand users and the wrong one at three million. The transition is hard, and the founder reluctance to make it is exactly what produces the kind of platform incident that erodes trust in a fintech.
Three readers, three different jobs to do.
Each card is a checklist for one role this conversation is most useful to. Pick the one that fits your week.
If you are a consumer-fintech founder
- Audit your funnel definition this week. If you are still counting sign-ups as acquisition, redefine acquisition as the first revenue-bearing transaction and re-baseline every dashboard against the harsher number.
- Map your retention engine and your acquisition engine onto two columns. If they are the same feature, the product has not yet found its second mechanic; identify the missing one before you raise the next round.
- Write down the cohort-retention curve at three, six and twelve months for your most recent batch. If the curve has not been published internally, the platform is running on faith rather than evidence.
- List every custodial responsibility your product currently holds and ask which could be moved to a regulated partner for a revenue share. The licence cost saved is almost always larger than the margin cost.
- Run one user-feedback cycle every two weeks at minimum. If your current cycle is longer than monthly, the iteration loop is slower than the cohort you are trying to retain.
If you build product for a young audience
- Treat your most adversarial users as the QA team. Build a triage process where every reported exploit is in the next sprint, not the next quarter.
- Find the over-adopted adjacent category your audience already lives inside. The right product surface is built from inside that category, not in opposition to it.
- Tokenise any flow where the user must trust the platform with money. The separation between user-facing currency and underlying asset is the cleanest design pattern for trust.
- Refuse the stealth-mode default. Ship a deliberately scrappy version on day one and replace it monthly. The feedback you cannot buy lives in the live app.
- Build your social mechanic before your monetisation mechanic. The cohort that brings a friend is worth ten of the cohort acquired through ad spend.
If you invest in early-stage consumer products
- Demand the cohort-retention curve before you read the sign-up chart. If the founder cannot produce the twelve-month curve, the platform is not yet evidence-driven.
- Read the definition of acquisition the company uses. If it is sign-ups or installs, ask why the harsher transaction-based definition has not been adopted. The answer is informative.
- Audit the partner stack. A two-year-old fintech with a clean partner-asset structure is cheaper to underwrite than one carrying custodial obligations beyond its operational maturity.
- Look at the team-shape decision. An interns-first early team can produce velocity; the question is whether the founders have planned the two-quarter window in which seniors arrive before the debt does.
- Ignore the marketing budget line and read the engineering investment in fault tolerance. In a money-movement product, the architecture is the moat, and the moat does not appear in the slide on customer acquisition cost.
The whole conversation, searchable and timestamped.
Search the text, highlight what matters, and export your highlights and notes together. Click any timestamp to jump in the video.