I own the numbers a business runs on — and I can build what produces them.
Chartered management accountant, which is a formal way of saying I am the person who asks where that figure came from — and who does not have to wait for someone else to go and fix it.
// the work
Work where the numbers had to hold up. Each one has a write-up — what was wrong, what I changed, and what I deliberately left alone.
Where a food producer’s numbers begin
The production system at a food manufacturer — every cost of sales, wastage and indirect-cost figure in the financials traces back to a run recorded in it.
Replaced a spreadsheet and a second round of manual re-keying into SAP. Five years in production, still running.
02Building a finance function from scratch
A growing health-tech startup with no finance function — no month-end, no management accounts, and a revenue number nobody could stand behind.
Revenue roughly 4x over two years, gross margin now stable enough to plan against, contribution positive for the first time — and a management pack the founder raised on.
03Roos Research
A subscription research desk for self-directed investors — turns market data and company news into a plain-language brief on a company.
Priced off its own per-analysis unit economics, and built so every figure is auditable — the model interprets numbers, it never calculates them.
// built for fun
Built because I wanted to, not because there was a business case. No revenue model, no roadmap — just things I enjoyed making.
// about
I'm a chartered management accountant. The training is in deciding what a business should measure and why it matters commercially — margins, unit economics, the handful of figures that actually tell you how a company is doing, and what it means when they move.
What I am good at is the bit in between. Most finance teams sit one step removed from their own numbers: waiting on IT for a change that takes a fortnight, keeping a shadow spreadsheet because nobody quite trusts the report, and finding three different answers to the same question. I close that gap — one agreed definition per number, produced the same way every month, fast enough to still be useful when a decision needs making.
This is not a coding course, and it is not a data engineering job with an accounting label on it. It is management accounting, done by someone who can build the thing himself — so the reporting matches how the business actually works, rather than how a ticket described it.
One rule runs through all of it, and it is an accountant’s instinct rather than an engineer’s: anything that can be computed gets computed in code, where it can be tested and re-run. Anything that cannot be checked does not get to decide. That is why the numbers still hold when somebody leans on them.
Based in Stellenbosch. The closest thing I have to a motto is that life's too short to braai kak wors — do it properly or don't bother. It turns out to apply to management reporting more than you would think.
- based
- Stellenbosch, ZA
- qualified
- chartered management accountant
- focus
- management reporting · forecasting · finance systems
- builds with
- python · sql · gcp
Got a number you can't trust, or a system that should be producing it?
A useful first message is one line about what you are trying to measure, or what currently takes too long to produce. I reply within a day.