Two findings shaped everything. The first was architectural. They asked us to solve consolidation, and the straightforward reading would have been to build a reporting layer on top of six systems. We proposed the opposite: build one unified system, so consolidation happens in real time and the problem simply disappears. The reason they had a consolidation problem was not that consolidation was hard. It was that they had six systems. Remove the six, and the problem is gone. They had asked us to solve the symptom. We proposed solving the structure. Most of what decides whether software works is decided before the build begins.

The second finding made the unified system buildable. The commercial side of all six companies was the same: purchase, sales, finance, payroll, compliance, reporting. What differed was production, because the group’s different food categories are made by genuinely different processes. So we built one commercial core and separate production modules, and linked each to the right companies. That is what let one system honor the real differences between the businesses without forcing them into a single mold.

And one detail shows what fit means here. Procurement in this trade carries five deduction factors on every purchase, including rain-affected damage, a category that exists because the harvest is exposed to the monsoon. No off-the-shelf system knows that category exists. We built it in as a first-class deduction, calculated automatically on every purchase. Fit is built, not bought.