Mariam Jimoh left a corporate career because she couldn't find the foods she grew up with in a mainstream UK supermarket — Ijebu gari, the right base for jollof, plantain that wasn't an afterthought stocked in a single dusty aisle. She didn't complain about the gap. She built the company that closed it.
Oja HQ launched in 2020 as digital grocery infrastructure for African and Caribbean communities across London, then Birmingham. By November 2021 she'd closed a $3.4 million pre-seed round led by LocalGlobe — an exceptional raise for a Black-founded consumer startup in a UK venture landscape where Black founders have historically received a fraction of a percent of total funding. She did it on a 200-user pilot with a 67% repeat-order rate, not years of revenue history. Raheem Sterling signed on as a brand ambassador. The community showed up in exactly the numbers the pitch deck promised.
The raise was never the question. The raise was proof the idea worked. What closed Oja wasn't the vision — it was what happened underneath the growth while nobody outside the company could see it.
What was happening underneath the growth
Building grocery tech for this market is not a simple e-commerce build, whatever it looks like from the outside. It requires real-time inventory sync across perishable stock with expiry-aware logic — the system has to know not just what's in stock but how long it stays good. It requires integrating suppliers who are often small, informal, and without API-ready systems of their own, meaning the tech has to bridge a real formality gap by hand. It requires last-mile delivery logistics that account for time windows and neighborhood access with real accuracy, payment infrastructure built for customers who may be underbanked, and what you'd have to call cultural inventory intelligence — knowing that egusi and plantain aren't a niche category, they're the product, and stockout logic has to be built around that instead of bolted on after.
That is a full-stack operational undertaking, attempted in an underfunded environment, for a community that deserved it, inside a venture landscape that historically hasn't valued either the market or the mission. And underneath the growth that funding was buying: eight legal claims for unpaid supplier debts, filed before the fundraising announcement even went public. HR run informally by a family member. A Q1 revenue dip with no runway built to absorb it. Salaries delayed. Vendors unpaid. None of that had to stay invisible until it became urgent.
This is the part of a founder autopsy that never makes the funding announcement: the eight legal claims, the informal HR, the missing runway math — none of it happened suddenly. Each piece was a reasonable decision made under real pressure, by someone building faster than any support system could keep pace with. That's the pattern I look for first in a Founder Autopsy, and it's almost never one catastrophic failure. It's a dozen small, defensible choices, made in isolation, that nobody was ever positioned to see stacking into a single unsustainable shape until the shape was already load-bearing.
The Questions I Would Have Asked Mariam
1️⃣ Was your team still playing telephone with suppliers about what was actually in stock, or did your inventory know in real time? Manually chasing that answer every week isn't operations — it's unpaid labor disguised as a spreadsheet.
2️⃣ Were the numbers you showed investors the whole picture, or the aggregate version that quietly hides which location or which order type is actually struggling? A number rolled up too far can hide the exact dip that needs the earliest warning.
3️⃣ When a stock item ran out after a customer had already paid, what happened next — a system that handled it automatically, or a person scrambling in a group chat while the customer waited to find out if she was getting a refund?
4️⃣ How much of your last-mile delivery ran on a system that actually thought for itself, versus somebody's gut call, made fresh, every single day?
5️⃣ Honestly — not aspirationally — how much leverage was your tech stack actually buying you per pound spent on it? Infrastructure that doesn't return time, margin, or peace of mind isn't running the business. It's running alongside it, as its own separate cost.
The Baddie Stack Prescription 💅🏾📱⌨️
📊 A real-time cash flow forecasting dashboard. Automated weekly P&L visibility broken down by location and order type, so a Q1 dip surfaces as a 30-day warning with time to act — not a shutdown trigger discovered after the fact, in an aggregate number that hid exactly where the problem started.
🏦 Vendor payment automation. Scheduled payments, supplier communication, and invoice tracking running outside the founder's own head and inbox, so eight legal claims for unpaid debts never has the chance to become the quiet, compounding story underneath a public fundraising announcement.
📈 Investor reporting automation. Recurring update flows and KPI dashboards that reduce the manual ask on the founder's time and keep investors genuinely engaged with the real numbers before a crisis — not scrambling to explain one after it's already public.
This breakdown can be customized with a Digital Systems Scan
Oja shut down in July 2023 after struggling to secure a follow-on round during that Q1 dip. Jimoh drew no salary for over a year and put her own savings into the business trying to hold it open. That's not a story about a founder who couldn't execute — the raise itself was proof the execution was real. It's a story about the follow-on funding architecture failing to show up exactly when the first round's thesis needed reinforcing, in a market that was never built to make that easy for her.
The uncomfortable question worth sitting with, whatever business you're actually running: if a stranger pulled your last twelve months of cash flow, supplier payments, and HR decisions into one document today, would it tell a story you'd recognize — or one that surprises you as much as it would surprise them? Most founders have never actually looked, because looking requires infrastructure that shows the whole picture at once, not four different half-updated spreadsheets and a gut feeling about how things are going.
There's a version of this worth holding onto, too, because it's easy to lose in a story this hard: the raise itself proved the market believed in her, and the community's repeat-order rate proved the customers did too. Nothing about what came after erases that. It just means belief and infrastructure are two different things, and only one of them fails quietly, out of view, until the numbers finally force the conversation.
That's the distinction worth carrying into any founder story that ends this way — including your own, if you're building something right now. The vision being real was never the question.
You cannot fund what you cannot see, and you cannot see what the infrastructure was never built to make visible. Mariam built something her community genuinely needed. The next version of this — hers or somebody else's — deserves the systems to survive what this market will throw at it.
Start with a Digital Systems Scan: Discover
Stay Savvy, Baddies 💅🏾.
Tech Baddie Out.