Recovery scenarios and uncertainty

Understand the constraint

Return to the common Northstar requirement of twenty nine good bottles. Supplier, manufacturing, and shipping ceilings are twenty four, thirty two, and thirty. Buying five additional on-time supplier equivalents for ten currency units each closes the nominal gap and costs fifty. Buying eight equivalents costs eighty but permits only thirty shipments because shipping becomes limiting. Against the twenty nine requirement, the extra three units beyond the needed five do not improve current due service. They might have later value, but that is a separate decision.

Now consider a downside case in which shipping falls to twenty seven. With the five-unit supplier recovery, completed flow is min(29,32,27), or twenty seven, leaving two late. A backup carrier can add three shipping slots for a fixed fee of eighteen, restoring the needed flow to twenty nine. An action should be checked against all downstream gates and should be charged according to its actual cost rule. Fixed fees differ from per-unit costs. Do not multiply a fixed pickup fee by the number of bottles.

Interpret the result

Without probabilities, compare scenarios without claiming an expected cost. If a decision maker supplies explicit probabilities, a weighted result can be calculated, but that does not guarantee a particular outcome. For teaching, use a sixty percent nominal case and forty percent shipping-downside case, and treat these probabilities as assumptions. Supplier recovery alone yields weighted shortfall of point eight bottle, although actual shortfall is either zero or two. Paying for contingent backup only in the downside has expected added fee seven point two, assuming activation is reliably available.

Create one row for each resource or stock position and one column for each time bucket. Put the unit in the row label and keep inputs separate from calculated results. Begin with the smallest case you can verify by hand, then extend the same equations across periods. Give every receipt a usable date and every requirement a due date. Preserve the baseline before changing an assumption so the effect of an action is visible.

Validate and explain

Use a second check that follows the physical story. Count what enters, what is used, what waits, and what leaves. A plan that balances in total can still fail at an earlier cutoff. Mark any unavailable input as an assumption rather than filling it with a convenient zero. A zero means a known absence; an unknown value means the promise remains conditional. This difference belongs in both the model and the decision note.

A calculation becomes useful when it changes a choice. Compare the baseline with at least two scenarios, identify the first limiting condition in each, and explain what additional resource would actually change delivered output. Keep a service measure alongside any cost or value measure. State whether the choice maximizes value, protects a commitment, or reduces the worst shortfall. A different goal can produce a different sensible answer from the same facts. Finish by testing one deliberately wrong proposal, such as a receipt used before arrival or a quantity above a resource ceiling. Your table should make that error visible. Save the rejected proposal with a short explanation, because it teaches a reviewer how the checks work. The eight hours include rebuilding, scenario work, writing, and teach-back; they are self-paced estimates rather than measured completion times. Faster readers should deepen the checks instead of merely rereading the text.

Build, test and explain a plan

Build a two-period scenario table. Each period requires 29 good bottles and has baseline ceilings supplier24/make32/ship30, with no carry. In period two downside shipping is 27. Compare no action; buy five supplier equivalents each period at 10 each; buy those plus a guaranteed backup reservation costing 6 upfront and 18 only if used. The reservation adds three slots in the downside. Assume downside probability 40% and the reservation is the only way to ensure access. The eight-hour estimate covers an original workbook or paper model, scenario calculations, checks, and communication; it is not experimentally validated.

Use paper or a spreadsheet. The estimates below include the study and worked example already on this page.

Your deliverables

• A reproducible planning table with inputs, formulas, units, and checks.
• A scenario comparison with service and resource consequences.
• A decision memo and short teach-back outline.

Keep these three ideas

• Evaluate recovery as a chain of contingent effects.
• Keep quantities, resource units, and due times explicit.
• Recheck all constraints after a change, and report remaining uncertainty.

Common mistakes

• Treating capacity in a later time bucket as available before the due gate.
• Calling an expected or conditional quantity a guaranteed shipment.