I built Owo in less than a day.
Not a prototype. Not a landing page. A functioning valuation tool for the Nigerian Stock Exchange — with a Supabase database, Vercel hosting, a live domain, and a data pipeline pulling sector betas and equity risk premiums from Damodaran’s annual dataset.
I’m a PM, not an engineer. Six months ago that sentence would have meant I needed to find a technical co-founder, write a spec, wait for a sprint, and ship something three months later that was 60% of what I had in my head.
Claude Code changed that equation.
The problem I saw
I spend time in Nigerian investment Facebook groups. What I kept seeing was the same pattern: someone posts a hot take on a stock — Dangote Cement, GTCO, Airtel Africa — and hundreds of people pile in with opinions, none of them grounded in any valuation.
It’s not because Nigerian retail investors are unsophisticated. It’s because nobody built the right tool for them.
Every global valuation tool — Bloomberg, Simply Wall St, Morningstar — uses US market assumptions. Risk-free rate based on US Treasuries. Beta from US market data. Equity risk premium calibrated to a market Nigerian investors don’t participate in. Run a Nigerian stock through those tools and the output is meaningless.
The data to do this properly exists. NGX filings are public. The CBN publishes the FGN bond yield. Damodaran releases his equity risk premium and sector beta data every January at NYU. Nobody had assembled it into a tool a normal Nigerian investor could actually use.
That was the problem. I decided to build the solution.
What I needed to build
Before writing a line of code — or asking Claude Code to — I did the product work myself.
The core insight: Owo had to be Nigeria-native by default. Not a global tool with Nigeria as an afterthought. Every parameter had to reflect the actual market:
- Risk-free rate: CBN FGN bond yield, not US Treasuries
- Equity risk premium: Nigeria-specific country risk premium from Damodaran’s emerging markets dataset
- Beta: Damodaran sector averages for emerging markets, not S&P 500 regression
Those aren’t configuration settings users toggle. They’re the defaults — already right when you open the tool.
I also made a deliberate model choice: start with DDM (Dividend Discount Model) because it requires the least data and covers the most important NGX stocks. DCF comes in V2 for Oil & Gas and Telecoms. Multiples screening after that. Sequence the build around what delivers value fastest, not what’s most technically impressive.
The other product decision that mattered: show the T-bill alternative on every output. At 22.52% risk-free return from CBN T-bills, every NGX stock has to clear a high bar. A stock showing a BUY signal at 8% dividend yield isn’t a buy if T-bills are paying 22%. That context belongs on every screen, not buried in a methodology page.
None of that is engineering. That’s product thinking. I did that part myself.
What Claude Code handled
Once the product decisions were made, I opened Claude Code and described what I needed to build.
What followed was the part that would normally take weeks:
Database setup on Supabase — schema design for stocks, prices, valuations, user watchlists, and the price_history table that will become Owo’s proprietary NGX archive over time. Claude Code scaffolded this, wrote the migrations, and connected it to the Next.js app.
Hosting on Vercel — deployment configuration, environment variables, build settings. Done.
Domain wiring — owongx.xyz connected, DNS configured, SSL sorted.
Damodaran data pipeline — this is the one I’m most proud of. Damodaran publishes his sector betas and equity risk premiums annually at NYU. The data is public but it’s in spreadsheets, not an API. Claude Code wrote the pipeline to pull, parse, and load the relevant emerging markets parameters into Supabase so Owo’s valuations stay current when Damodaran updates his data each January.
The valuation engine — DDM implementation with the Nigeria-specific parameters I’d specified. Cost of equity calculated from CBN Rf + (beta × ERP). Fair value derived. BUY / HOLD / SELL verdict generated against current market price.
Email triggers — automated digest for watchlist alerts when a stock’s verdict flips.
Every one of those tasks is real engineering work. Each one would normally require me to either learn it myself over days or hand it to someone else and lose control of the output. Claude Code did them while I stayed in product mode — thinking about the user, the model, the market, the business.
What I still had to be the human for
Claude Code is not a PM. It does not know that DDM is structurally unreliable for high-ERP markets. It does not know that Nigerian investors need to see T-bill opportunity cost next to every valuation output. It does not know that the real value proposition is historical — showing investors when stocks were cheap, not just whether they’re cheap today.
Those were my calls. Every one of them.
The product strategy — who Owo is for, what problem it solves, how it’s differentiated from every global tool that treats Nigeria as an afterthought — that’s not something you can delegate to a coding agent. The freemium model, the sequencing of DDM before DCF before multiples, the decision to build the price_history archive now even though its value compounds over 12 months — those required judgment about the market, the user, and the business.
Claude Code is extraordinarily good at closing the gap between a decision and its implementation. It is not good at making the decision.
That division of labor is the thing worth understanding.
What this means for PMs who can think but couldn’t build
There is a class of product manager who has always been able to think clearly about problems, users, and strategy — but who needed an engineer to translate that thinking into something real. The cost of that dependency is significant: slower iteration, diluted vision, ideas that never get tested because the queue is too long.
Claude Code does not eliminate engineering. For complex systems at scale, you still need engineers. But it dramatically lowers the threshold for what a non-engineer can ship independently — and it does it fastest when the person using it has strong product judgment.
The Supabase schema Claude Code wrote for Owo is good because I knew what data the product needed and why. The valuation engine is accurate because I understood which parameters mattered and made the right choices about Nigerian market defaults before the first line of code was written. The email pipeline makes sense because I knew what user behaviour I was trying to drive.
Garbage in, garbage out — but at ten times the speed in both directions.
Where Owo is now
Owo is live at owongx.xyz. It covers NGX-listed stocks with DDM valuations, daily price tracking, T-bill comparison on every output, watchlist functionality, and email alerts when verdicts flip.
The price_history table started accumulating in May 2026. In twelve months it will be a proprietary NGX financial archive — daily fair value and verdict data for 186 stocks that no competitor can replicate without starting from scratch. That’s the moat, and it’s compounding already.
DCF for Oil & Gas and Telecoms is next. Then multiples screening. Then the historical intelligence layer — the feature that answers not just “is this stock cheap today” but “when was this stock last cheap, and are those conditions approaching again.”
The GTCO investors who bought at ₦66 made 136% in twelve months. The ones who bought at ₦157 are holding an overvalued position. The difference between those two investors is not intelligence. It is information and timing.
That’s what Owo is for.