Process Notes — Issue #008 — Scale-Up Playbook

Pilot-to-Production: The Knowledge That Didn’t Transfer

Why a complete handover document isn’t the same as a complete knowledge transfer

The Wrong Suspects

The investigation followed its predictable path. Formula reviewed. Raw material CoAs pulled. Micro results checked. All clean.

This is where most scale-up failure investigations stall — because the instinct is to examine what was made, not how the process transferred. The product specification was fine. The process rationale had never been written down.

The distinction matters more than most transfer documents acknowledge. A specification tells production what the output should look like. Process rationale tells them why each parameter was set where it was, and what happens when it drifts. Production had received the former. They were troubleshooting a process they had only been given targets for — not understanding of.

What the Pilot Data Actually Captured

Pilot scale is a managed environment. During development, the R&D team compensated for process drift in real time — extending mix time when viscosity looked low, adjusting jacket temperature manually when come-up was slow, reducing fill rate to hit the texture window. None of this was logged as critical. It felt like normal adjustment, not transferable process knowledge.

At production scale, those adjustments were no longer available — not because operators were less capable, but because nobody had told them the adjustments existed, or why they were needed.

Five parameters consistently carry this risk in scale-up:

None of these transfer linearly. A vessel twice the size does not transfer heat twice as fast. An agitator running at the same RPM in a larger vessel delivers a fundamentally different shear profile — tip speed may be identical, but bulk flow patterns change with vessel geometry. A centrifugal pump operating at higher throughput applies different shear to a sensitive product than the peristaltic it replaced at pilot.

At pilot, the R&D team absorbed these differences intuitively. At production scale, they were invisible — because they had never been made explicit.

The Gap Revealed

What the production team received: a recipe, a process flowsheet, an in-process specification sheet. What they did not receive: the reasoning behind each critical parameter, the range within which the product remains stable, and the in-process signals that indicate drift before a batch goes out of specification.

The transfer document was complete on paper. It was empty of process understanding.

A Technical Transfer Is a Knowledge Transfer

This distinction reframes what a transfer document is actually for. Its purpose is not to record what was done. It is to give production the ability to understand, run, and troubleshoot the process independently — without the R&D team in the building.

Critical Process Parameters vs. Non-Critical Parameters

Not every process variable carries the same risk. Critical Process Parameters (CPPs) are those whose variation directly impacts a Critical Quality Attribute — texture, viscosity, microbial load, fill weight. Non-critical parameters can drift within a wide range without affecting output quality.

The problem is that this distinction is rarely made explicit in transfer documents. Everything gets listed at the same level of importance, which means production treats everything with equal attention — or none at all.

CPP identification should happen during development, not after the first production complaint. By the time you’re troubleshooting week three, the decision about which parameters are critical has already been made for you — by the failure.

The Process Design Space

A single-point specification — mix for 8 minutes at 72°C — tells an operator what to aim for. A process design space tells them what happens if they miss, and by how much they can miss before the product is at risk.

The design space is the proven range within which each CPP can operate and still produce an in-specification product. It is established during development through deliberate variation studies — not assumed from a single successful trial.

When production has the design space, an operator who sees temperature drift to 69°C knows they are still within the safe operating range. Without it, they have no reference point — and either over-react to normal process variation or miss a genuine deviation entirely.

The Transfer Meeting Is Not the Transfer

Verbal handovers decay within weeks. The R&D engineer who ran every pilot batch is not present on week three of production when something drifts at 6am on a Monday. The document must carry the understanding independently. If it cannot be used to troubleshoot without the author in the room, it is incomplete.

Process Notes — The weekly digest for food manufacturing professionals

processnotes.beehiiv.com