The brief had a hard constraint attached: the line does not stop. Whatever we did to the system running this factory, production could not pause for it.
What we cut
The existing ERP had accumulated fifteen years of features, and most of them were unused. We spent the first week watching people work rather than reading the spec, and found the daily reality ran on about a fifth of the system.
So we rebuilt that fifth properly and left the rest running on the old system until it could be retired quietly.
What shipped first
Inventory, because it was the thing costing them money every day. Getting one painful area right and in production beat a complete system six months later.
What we kept
Their data model, largely. It was unfashionable and it was correct — it described their factory accurately. Rewriting it would have been us preferring our own habits to fifteen years of their hard-won knowledge.
The whole thing shipped in six weeks, with a documented API, and the line never stopped.
Written by Aquib ul Haq
Talk to us about your project→