What actually transfers between markets, what doesn’t, and how the data you gather along the way becomes the thing competitors can’t copy.

Expansion is usually described as expensive and slow, and it is when you rebuild from zero each time. The alternative is to work out what genuinely carries across, then port that.

We asked Orca Fraud co-founder and CEO Thalia Pillay, whose platform now screens around $5bn a month across more than 70 countries, how they did it…

Orca’s founding hypothesis was that fraud in Africa looks different: different payment methods, mobile wallets rather than cards, different socioeconomic behaviours underneath the transactions. That turned out to be true about the surface and not about the machinery, which is exactly what makes expansion fast. The underlying models travel; the local specifics get rebuilt.

“We built our moat on data and governance.”

They didn’t assume African fraud was different; they went and checked: roughly 150 conversations with fraud analysts across the continent, the people who see the problem daily. That scale of interviewing is unusual and it’s the whole foundation, because it turned a plausible idea into something they knew. If your expansion thesis rests on a belief nobody has tested with practitioners, that’s the cheapest thing to fix first.

The reusable asset is rarely the finished product; it’s the layer underneath. Orca reuses its machine learning models across emerging markets, which is how M-Pesa fraud monitoring in Kenya became Pix monitoring in Brazil almost overnight, then PayShap monitoring with South African fintechs. Different rails, different countries, same underlying engine. Be precise about which part of your product is the engine and which part is local plumbing.

Reuse only works if the markets rhyme. Orca deliberately focuses on the Global South and emerging markets, where the payment behaviours and infrastructure share enough DNA for the models to carry. Picking a second market because it’s large or fashionable, rather than because it resembles the one you understand, is how you end up starting from scratch after all.

The unglamorous constraint is integration. API maturity across the continent is low, so Orca had to innovate on deployment rather than insist customers meet a standard. In Thalia’s words, many customers send them data via osmosis and love a spreadsheet, and the job is taking them from a skateboard to a Ferrari rather than refusing to work with the skateboard. Expansion dies on the assumption that a new market will integrate the way your first one did.

Each market you enter produces data nobody else has, and across enough markets that becomes genuinely hard to copy. It de-risks the business too, since no single country can sink you, and it makes a far stronger investor story. Orca found niches in markets like Ghana and Uganda, and Thalia’s argument for looking beyond South Africa is blunt: you’d rather capture 10% of a large market than 10% of a micro one.

Done this way, each new market costs a fraction of the first and leaves you with something a competitor can’t buy: a data set built from countries nobody else bothered to serve properly. Worth saying it’s not the only sane answer, though. At the same event, Zimi’s Michael Maas argued the opposite case for hardware, where crossing a border means new compliance, new partners, and physically shipping units, and going deeper at home wins. Which camp you’re in depends on how much of your product actually travels.

Thalia’s full talk from Founder Collab Live is available to members inside the Founder Collab, where she goes further than we could cover here:

You’ll also get access to 40+ masterclasses from SA founders and operators on sales, fundraising, UX, paid media and more inside The Founder Collab.

Get more SA tech and business news, tips and business-building workflows, and subscribe to The Open Letter.