The shortcuts stopped scaling
At Foody, a strong engineer could carry a surprising amount of operational context in their head. It felt fast because coordination was informal. Inside a global organisation, the same habit became fragile.
Scale exposes every ambiguous boundary: who owns a service, which failure matters first, how a decision is recorded, and what another team can safely assume. More process is not automatically the answer. The useful work is designing interfaces between people as carefully as interfaces between systems.
My job shifted from solving the problem in front of us to helping many teams solve related problems without waiting for the same few people.
That meant making priorities clear, recording decisions and reducing the amount of coordination needed to move. The work had to depend less on any one person making the decisive call.
The market did not become generic
Global organisations benefit from common platforms, but markets remain stubbornly specific. Payment habits, merchant operations, geography and customer expectations do not become uniform because the organisation chart does.
Calling the tension central versus local turns it into a political contest. A more useful question is which decisions benefit from shared infrastructure and which still need direct market knowledge.
Standardisation works when it removes repeated undifferentiated work. It fails when it erases the information that made the local product useful.
Speed had acquired a cost
The startup had been fast partly because some safeguards did not exist yet. A large company sees failure modes that a small one can avoid only because it has not reached them. Not every extra control is bureaucracy, and not every control is useful. Working inside both made the difference easier to see.
Public record
These sources confirm the public facts. The rest is my recollection, written in 2026.