Record Research Inc.
Decades of Billboard chart data moved off SQL Server 7, Classic ASP and Visual FoxPro onto Azure, while the books the business sells kept getting published.
All work · Media & entertainment
Seventeen years through one consultancy. At UMG, a catalogue feed reconciled field by field against the rights database, and the slowest queries in the business cut from thirty seconds to under one.
A catalogue data dump had to be reconciled against the internal rights database, and the feed had to be specified well enough that outside organisations could build against it without seeing UMG's internals. Separately, the core queries took thirty seconds each.
Reconciled the dump field by field and wrote the feed specification as two deliberately separated documents: a functional description with internal processing stripped out, and an implementation document written so it could be handed to any outside organisation. Later, for the same consultancy, architected a festival ticketing and cashless-payment system: scheduled cloud jobs pulling events and venues from three ticketing platforms, with a point-of-sale for festival vendors charging a customer's in-app wallet. The requirements existed only as PowerPoints and mock-ups that contradicted each other, so I refused to start from screens and got the process agreed first.
Core query times from thirty seconds to under one. The only client relationship in the archive still live at its end.
Stack
Tell me what it does for the business, what it is written in, and what stops when it goes down. I will say plainly whether I can help.
Tell me what it runs on