Who Runs Your Schedule When the Scheduler Is Gone

A senior scheduler resigning costs you far more than a planner. The logic that decides which suite absorbs a delay, how long a changeover really takes on a bad day, and which campaign can be broken without wrecking the week is almost never assembled in one place, and whatever is left of the notice period is how long you have to pull it together.

Backfilling that person is harder than it used to be. A Deloitte study, reported by PharmaVoice in October 2026, sets out how thin the hiring pool has become.

Manufacturing operations and quality are among the hardest roles for biomanufacturing sites to fill, which means a replacement planner may not arrive inside the notice period at all.

Most scheduling rules are written down somewhere already

Most of what governs a schedule is written down somewhere. A clean hold window will be in a validation report, while a campaign minimum often traces back to an S&OP decision nobody has revisited since. Other rules sit in procedures saved as PDFs that only a few people know the location of.

The difficulty is assembling all of them at once. Nothing in a batch record tells the next planner that line 3 needs an extra forty minutes when it follows the high-potency product. Nor does anything explain that a particular campaign can run eleven batches even though the written rule says nine. Some of that sits in a document, others sits in a spreadsheet one planner maintains, while the rest sits in a head, and none of it is indexed against the others.

Sites that do this well have made the aggregation somebody's job, but, most have not, and they discover the gap in the week they can least afford to.

One strong scheduler looks exactly like good scheduling

A site running on a single strong planner shows none of the symptoms you would think to look for. OTIF holds and disruptions get absorbed, so nobody escalates because nothing goes wrong often enough to escalate. The dependency stays invisible right up until the week the resignation lands, which is the point at which there is nothing to be done about it.

What the new planner inherits

Watch what gets asked in those sessions. Somebody wants to know where the current spreadsheet lives, and who to call when lab results are late. All of it is answerable and is written down in an afternoon. Nobody asks why the coating campaign always runs before the high-potency product, because nobody outside that planner's head knows it is a rule rather than a habit, and the question does not occur to the person leaving either.

So the workbook gets passed on and the reasoning stays put, and the new planner inherits numbers whose origin they cannot reconstruct.

Formal training runs around three months, and competence takes considerably longer, because the rest is learned by navigating the systems and finding out which sequences generate deviations, including the ones that embed a mistake and cost a line stop to unpick. Through that whole period the new planner is working to produce one plan with no errors in it, rather than a plan that maximises output. The schedule stays compliant and the plant runs below what it could, which never appears as a deviation and never shows up in anybody's report.

Documentation and cross-training treat the symptom

Most sites reach for documentation first. Asking a departing scheduler to write down what they know produces a document that is always thinner than the knowledge, because people cannot list rules they apply without noticing they are applying them. What reaches the page is the part that was already explicit.

Cross-training helps without solving it, since two planners holding overlapping mental models that neither has written down just means two departures to worry about instead of one.

Neither approach is a bad idea; they both treat the symptom, which is that a person is leaving, rather than the actual cause, which is that the plant's scheduling rules were never assembled into one reviewable artefact.

Getting the model out of the person's head

Get the constraint model out of the person and into something a second person can read.

Start by listing what actually binds the schedule, in the planner's words rather than the SOP's. Hold times and cleaning windows come first, then the rules about what can follow what, then who is qualified to run it. Once that list exists, mark every rule nobody can source. Those are the ones that leave with them, and they are the reason the next planner rebuilds the logic by making the same mistakes.

The benefit shows up long before anyone resigns. A written model is reviewable, so it gets corrected when it is wrong instead of quietly compounding. It can be handed over in a week rather than a year. And it survives its author, which is the only real test of whether your scheduling function is a capability or a person.

Alexandre Clarizio is founder and CEO of Kaster Technologies, which builds AIM, a decision intelligence platform for GMP pharmaceutical manufacturing.