Case Studies · Deluxe, studio-to-streamer delivery
Every vendor ran its own QC, and the same defects kept recurring
45%
fewer QC failures across vendors
Systemic change
A systemic fix across an ecosystem of partners, not just the internal team.
The situation
The delivery pipeline from a major studio to a global streaming platform ran through several partners, each responsible for one part of the chain: encoding, localization, packaging and delivery. Each partner had its own QC tooling and its own definition of a pass. The result was predictable. A file could clear one partner's check and fail the next, or clear every check and still fail at the platform.
The same categories of defect recurred month after month, rework was eroding margin, and launch dates were slipping. Separately, asset retrieval was slow because storage was scattered across systems with inconsistent metadata, so finding the right version of the right file was its own small project.
The cause
There was no shared definition of quality. Each partner had been onboarded independently, with its own checks accepted as-is, and nobody had reconciled them. Defects were being measured by partner rather than by category across the whole chain, so the pattern was invisible to everyone who only saw their own piece.
Action
I designed a centralized QC framework: a common set of technical standards, automated checks for the measurable items, and a single validation logic, so that a pass meant the same thing at every point in the chain. That was the easier part. The harder part was adoption. I worked with each vendor in turn to map their existing checks onto the shared standard, close the gaps, and run the same validation on their side before handoff.
In parallel, I led the move to centralized asset storage with consistent naming conventions and a standardized metadata schema, and trained teams on tagging so that retrieval depended on the system rather than on someone's memory.
The result
QC failures across the vendor network dropped by 45 percent, and rework costs fell with them. Asset retrieval time was cut roughly in half, and delivery success rates to the platform improved. The framework became the shared baseline for the account rather than an internal standard.
Defects addressed
- —Vendors were onboarded without a shared definition of done. One written standard now exists, and every partner added after it inherits it.
- —Metadata and naming standards lived in a wiki. They now live in the vendor agreement.
- —Defects were measured by partner. They are now measured by category across the whole chain, monthly.
Lessons learned
Before a second partner joins any pipeline, one written definition of done should exist, and it should be part of the agreement rather than a reference document. Defect data should be tracked by category across the entire chain so recurring patterns surface as standards problems, and the standard itself should be reviewed on a schedule rather than after the next failure.
This is the kind of situation I consult on under Program Management.
