Work
Fintech / Payments / Edtech

Who gets to control a child’s money?

Earlybean began as a financial-literacy product for children and evolved into a family-finance and school-payments ecosystem. I co-founded the company and led product direction and hands-on design across the parent, child, school and merchant experiences.

Role
Co-founder & Product Lead
Timeline
Jun 2020 — May 2026
Company
Earlybean · Techstars ’23
Four actors — parents, children, schools and merchants — each with the thing they asked for, beside a panel stating what the product had to become: the parent as the authority over a child account rather than a customer of a new bank.
One child’s money, and the four parties with a claim on how it behaves.
01

A financial-literacy app was not enough

Earlybean started with lessons, savings goals and a child-centred experience. It taught the mechanics of saving, but gave families limited ways to apply them with real money.

02

What parents and schools actually needed

Across school conversations and repeated parent feedback, the same tensions kept surfacing. Parents did not want another standalone banking relationship: they wanted to fund children from accounts they already used, retain control, and understand how money was being used. Schools wanted oversight inside their own environments. Children needed safe, practical places to save, spend and transfer real money.

Each of those tensions belonged to a different party, and none of them could be resolved inside a child-facing app. That is what turned a product question into an architecture question.

The four actors, live: click each to read what they asked for and what it forced the architecture to do.
03

Reframing the product around relationships

The product stopped being only a child-facing app. It became a system connecting a parent’s authority, a child’s bounded financial identity, a school’s oversight and operating role, and a merchant’s ability to receive and manage money. Trust, permissions and money movement became the architecture; the interfaces followed from that model.

Those three relationships do not run in the same direction. Money enters from a parent’s existing bank and leaves to an approved merchant or a validated account; authority runs from the parent over the child’s controls; oversight runs from the school across its own environment and no further. Reading them as one diagram is what makes the boundary visible.

The four-actor model, live: switch between money, authority and oversight. The dotted boundary is the reason direct spending stays inside approved environments.
Over the child’s moneyParentChildSchoolMerchant
Fund the accountYes, from an existing bank accountNoNoNo
Set limits and block spendingYes, per railNoNoNo
Approve a request or release a reached goalYesCan requestNoNo
Spend by card or tagNoWithin limits, approved environments onlyNoReceives, cannot initiate
Transfer to an Earlybean ID or bank accountYesWithin limitsNoNo
See activityEverythingOwn balance, goals and requestsIts own environment, per linked studentIts own takings
Approve which merchants a tag works atNoNoYes, inside the schoolApplies
Who may do what over the same money. The matrix was settled before any screen: every screen in the case is one row of it, seen from one column.
04

Giving parents authority without another banking app

Parents needed distinct controls for funding, approvals, limits, visibility and blocking. The product treated the parent as the authority over a child account rather than a customer expected to adopt a new bank, so parents could fund children from accounts they already used while retaining control over what happened next.

Keeping those five controls separate is the whole design. Collapsing them into one “parental controls” screen is what makes products in this space unusable: a blocking switch and a spending limit are different decisions, taken at different moments, and a request from a child is not a notification to be dismissed.

The parent app, live: toggle the controls and switch child. The parent is the authority over an account, not a customer being asked to adopt a new bank. Redrawn from the shipped product with representative data.
The child app, live: switch the three tabs to see what each surface is allowed to contain. The child owns a real identity — a tag, a balance, a request they can make — inside limits someone else set. Redrawn from the shipped product with representative data.
05

Making money behave differently by purpose

Money in a savings goal was locked and not spendable. Money loaded onto a card or wristband could be spent with approved merchants inside school environments or curated events. Transfers ran on a separate rail through Earlybean IDs or validated bank accounts. The product had to make those states and rails legible because each carried different risks, permissions and expectations.

The three rails, live: one balance would have been a lie about what the money can do. Each rail carries different permissions and risks, so the interface shows the difference rather than a single number.
Savings goals, live: step through saving, nearly there, reached and waiting on a parent, and released. The four states are the whole design problem — a reached goal is an authority decision, not an unlock. Redrawn from the shipped product with representative data.
06

Turning schools into safe financial environments

Schools needed more than a fee-payment endpoint. The product direction covered onboarding, monitoring, school payments, ancillary payments and oversight across school-owned and third-party merchants. Direct spending stayed restricted to approved school environments, giving schools control over merchant acceptance and reducing the risk around children carrying spendable money.

School admin, live: a school needs to know three things per student — who can authorise, whether the tag works, and whether anything unusual happened this week — so those are the three columns. Names, guardians and tag numbers are masked. Redrawn from the shipped product with representative data.
Merchant acceptance, live: click a place to follow the chain that decides a tap, what the child reads when it fails, and whether that route shipped.
07

The tap, and the decline that explains itself

Everything else in the system exists to serve one moment: a child taps at a canteen till with a queue behind them. A decline there cannot be an error code. It has to name the rule that stopped it and who can lift it, because a child at a till cannot debug a permission model, and a short school-break rush is no time to be working one out.

The canteen till, live: build a basket and try the tap, then step through the four results. Each decline names the rule and who can lift it. The menu, the prices and the child are drawn for the frame. Redrawn from the shipped product with representative data.
08

What shipped, what remained directional

Shipped: parent-controlled child accounts, savings goals, card and wristband spending, Earlybean ID and bank transfers, school onboarding and monitoring, and the merchant application. Directional, and not presented as shipped: advanced school messaging, the complete school money OS, deeper school-owned and third-party merchant management, revenue-share or rental tracking, and the broader ancillary-payment operating model.

The shipped and directional split, live: click any item for what it is and why it landed there. The platform’s figures stay in the written case study below, and no frame in this set carries one.
09

Evidence and critique

3,416
Active users
₦48.29m
Cumulative inflows
₦43.61m
Cumulative outflows
10
Schools and merchants onboarded

Backend snapshot dated July 2026: 3,416 active users across parent and child profiles; ₦48.29m in cumulative inflows; ₦43.61m in cumulative outflows; and 10 schools and merchants onboarded. These are separate platform measures. Inflows and outflows are reported as recorded and have not been netted.

The snapshot proves production use, active accounts and real money movement. It does not tell us how often families returned, how activity changed over time, or how deeply individual schools adopted the platform. The broader school operating model also remained only partly implemented.

FintechPaymentsEdtechCo-founderTrust design
Next case study
The customer journey ends in a kitchen.