Two graduates in Lima want to start a fintech. Their first idea dies on contact with the market — almost every adult in Peru already pays with one app. Over three sessions you will do what they have to do: find an opening an incumbent can't close, and defend it out loud. You will leave with a one-page pitch and a test you could run on Monday.
How to use this pack. This is the case — read Part 1 before Day 1, because class time is for building, not reading. How the three sessions run, what your role is, and what you hand in are all in the workshop guide, which is the same for all four workshops. Nothing in this pack is submitted. Daniela, Marco and their venture are invented, a realistic teaching device; everything about the market is real and drawn from public data (sources at the end). This workshop is Module 1 of Global Financial Innovation, a COIL course taught between Lexington and Lima.
The market, the incumbent, and the method. Roughly 20 minutes of reading. Come to Day 1 having done it — the room moves fast.
Daniela Quispe grew up behind a fruit stall in one of Lima's big wholesale markets, helping her mother count change and chase down customers who paid "next week." Marco Salas is the friend who learned to code building inventory spreadsheets for that same stall. Fresh out of university, they want to start a fintech together.
Their first instinct is the obvious one: a payments app. A slicker, friendlier way to send money. They sketch it for a weekend. Then Daniela's cousin laughs at them: "Everybody already has Yape. Why would I download yours?"
She's right, and the numbers prove it. In a country of roughly 34 million people, one app — Yape — is used by more than 20 million. Building "a better wallet" is like opening a new postal service when everyone already has the same one and it's free. The first idea is dead before the first line of code.
So Daniela and Marco face the real question, the one this workshop is built around: in a market an incumbent already owns, where is the opening — and how do you build something different enough to matter?
Daniela and Marco aren't crazy to want in. Peru's fintech scene is one of the most active in the region, and Lima is where nearly all of it happens. But "active" cuts both ways: a crowded field means more competitors fighting over the same easy ideas.
Two facts here will matter later. First, lending is the busiest segment but also the deepest need — a clue, not a closed door. Second, open finance (rules letting customers share their banking data with apps they trust) is arriving, which can hand a newcomer data it never had before.
Peru went almost overnight from a deeply cash-based economy to a tap-and-pay one. In 2018 the average Peruvian made about 25 electronic retail payments a year; by 2022 it was around 99 — the fastest jump in Latin America — and by 2024 the central bank counted roughly 442 digital payments per adult, more than one a day.
Two home-grown apps drove that. Yape, launched in 2017 by BCP (the country's biggest bank), and Plin, built by a group of rival banks. Together they cover most of the country and, since 2023, they even work with each other.
This is the wall, and there's a name for it: the network effect. A payment app is only as useful as the number of people you can pay with it. Yape is valuable because everyone has it — and everyone has it because it's valuable. A brand-new wallet, however beautiful, starts at zero people, so it's useful to no one, so no one joins. The product can be better and still lose.
The lesson generalises far beyond Peru: you almost never beat a network-effects incumbent by doing the same thing slightly better. You have to change the game.
An incumbent that serves "everyone" is optimised for the average customer and the simplest job: moving small amounts between people. That leaves whole groups under-served and whole needs barely touched. These cracks are where a newcomer looks — and on Day 1 you will be asked to make the case that each of these numbers is not a problem but a business.
Read those four numbers as jobs the wallets don't really do. Yape will happily move S/50 between two friends. It is far less good at lending a market vendor S/2,000 of working capital on Monday to repay on Friday, at helping someone with lumpy informal income actually save, at reaching a customer with no smartphone in a highland town, or at underwriting a tiny business the banks have written off.
Notice the pattern: the giants compete to move money. The open ground is in helping people put money to work — borrow it, save it, smooth it, reach it — for customers the mainstream ignores.
"Have a great idea" is not a plan. Good founders don't wait for lightning; they run a process. This is the one you will run on Day 1. The order matters — start with the customer, not the technology.
Resist "all Peruvians" — that's the incumbent's turf and far too broad to design for. Choose a specific group whose needs are poorly met today. This is your beachhead: small enough to win, real enough to grow from.
People don't want a financial product; they want a job done in their life. Look for a job that is frequent, painful, and currently handled badly — by a relative, by a loan shark, or not at all. Skip jobs already solved: "send money to a friend" is taken.
Ask what you can do that the incumbent can't or won't. Data they don't have, a customer they ignore, a channel they can't reach, trust they can't buy. Without a wedge, the giant simply copies you and wins.
Daniela and Marco run the method against the cracks in Exhibit C and reach four candidates. Each targets a different gap; each has a genuine catch. On Day 1 you may take one of these, or — better — find your own. There is no obviously correct answer, only better and worse arguments.
Picking an idea is half the work. The other half is making sure it's differentiated — positioned somewhere the incumbent isn't, on a dimension customers actually care about. "Slightly cheaper" or "slightly nicer" is not differentiation; it's an invitation for a 20-million-user giant to copy you next quarter.
A positioning map makes this visible. Put the choices customers care about on two axes, plot the incumbents, and look for the empty space. Below, the horizontal axis is who you serve and the vertical is what you do. You will plot your own idea on this map, on a whiteboard, in front of the room — so make sure you can defend the square you land in.
The orange product isn't "a better Yape." It sits in a different quadrant entirely: a narrow segment (market vendors) and a different job (credit, not payments). That distance is the whole point — it's hard for an incumbent built around the opposite corner to chase you there without breaking its own model.
The same logic applies to the others: B differentiates on behaviour and emotion, C on physical distribution, D on a specific flow. In every case the test is one blunt question, and it is the question you will be asked repeatedly in this workshop: "When the incumbent sees this working, why can't they just add a button and crush us next quarter?" If you can't answer that, you don't have a product. You have a feature waiting to be copied.
Three sessions, the same shape as every other workshop this semester. The full block-by-block timing, the four roles, and the deadline are in one place — read it once and it covers all four workshops.
This pack is the case — the market, the wall, the method and the four directions. The workshop guide is the process — who does what in each block, when the deadline is, and what your one-page brief must contain. Nothing in this pack is submitted.
Kill the obvious idea, narrow the customer three times, name the job and your first advantage, take the Skeptic’s attack, and commit. The idea is fixed at the end of this session.
Evidence round, rebuild, cross-examination between teams, then this workshop’s extra question — below.
One page, written by you, about your role. Posted in the team folder where your teammates can read it. This is the only thing you hand in.
Four briefs on the table, name the one thing that would kill the idea, design a cheap test, and pitch for three minutes. The pitch is spoken only.
Every workshop asks one extra question at this point, and it is the only block that differs between the four. For Payments it is:
What does one customer pay you in a year, and what does the incumbent charge for the nearest thing today?
A good answer is two numbers with sources. If your idea involves lending, give the loan size, the term, and what happens when one borrower in ten does not repay. If you cannot find a number, say so and say where you looked — that scores higher than inventing one.
Makes the call when the team is split, and presents the pitch on Day 3. Writes the decision memo.
Owns the evidence and the arithmetic. Asks where every number came from. Writes the numbers memo.
Holds one named person in mind and asks whether she would really use this. Writes the portrait.
Plays both outsiders. As the incumbent: “we’ll ship that as a button next quarter.” As the regulator: “which licence is that, and has the SBS seen it?” Writes the attack memo.
Between Day 1 and Day 2, everyone brings three facts with sources — one that helps the idea, one that hurts it, one that surprised you. Peru students should get at least one of them by talking to a real person: a market vendor, a bodega owner, a moto-taxi driver. US students should get theirs from sources and by pricing the incumbent — download the app, find the actual fee. The one that hurts is the one the class wants to hear.
Read the workshop guide — block-by-block timing, the four briefs, and the deadline table →
Peru is one answer to a universal question — "how does a cash economy go digital, and who gets left out?" Three other countries answered it differently, and the answer changes which startups are even possible. You'll be asked about this in the final debrief.
The central bank built a free, instant payment rail and required banks to join. Top-down and universal — so a startup competes on services built on top of the rail, never on the rail itself.
A shared public "plumbing" layer any app can plug into. Dozens of apps compete fiercely on experience because none of them owns the network — the opposite of Yape's private one.
Money went digital through a telco and a vast network of human agents in shops — reaching people with no bank and no smartphone. The closest model to Direction C.
Start here for the between-session evidence assignment. Every figure in Exhibits A–C traces to one of these.