Master planning runtime, stale safety stock and inventory that should not be replenished
Vantage Cycle Works builds bicycle frames in two plants and buys 3,200 finished goods and components from 190 suppliers. Average purchased-component lead time is 55 days, and 40% of annual volume ships from March through June. Planners run the deprecated master planning engine in a nightly batch that takes about seven hours; because they do not trust the output they rerun it three times a day and firm orders by hand. Safety stock is calculated in Excel and keyed into item coverage; nobody has updated it in 14 months. Expedited freight was 4.1% of purchased spend last year. Analysis shows 22% of SKUs had no demand in the last twelve months yet still carry roughly three months of stock, while seven high-runners caused 61% of last season's missed ship dates. The company owns Planning Optimization licences but has not enabled them, citing 'missing features' that nobody can name.
1. Should the company enable Planning Optimization? What must be validated before it becomes the only planning engine?
Read the worked answer
Enable it, but validate in a sandbox copy of production first, because the objection is vague and the current state is worse: a seven-hour nightly run that planners rerun three times a day is not a plan anyone trusts, and each rerun competes with other batch work.
What to validate:
- Feature fit against the current documented list of unsupported scenarios. Convert 'missing features' into named items. The list has historically excluded areas such as process-manufacturing coverage with formulas and co-products or by-products, so confirm whether any of the two plants' items use those patterns. If nothing on the list applies, the objection collapses.
- Output equivalence on a sample. Take 100 items across the seven high-runners and 20 slow movers, run regeneration on the sandbox, and compare planned orders, dates and quantities with the legacy result. Differences are either bugs or the stale safety stock being corrected; classify each one.
- Operating model. Planning Optimization runs as a service outside the batch server, so schedule regeneration and net change deliberately, decide who owns the planning workspaces, and stop the habit of daily reruns.
Keep the legacy engine available for one cycle as a fallback, with a named decision date.
2. How would you rebuild the safety stock approach so it reduces both expedite freight and excess stock?
Read the worked answer
Today safety stock is a single stale Excel number spread across 3,200 items, which explains both symptoms: too much where nothing ships, too little where the seven high-runners are. Rebuild on segmentation.
- Classify by demand volume and variability (for example high-volume stable, high-volume volatile, low-volume, inactive). Use the last twelve months of demand and the 4.1% expedite spend by item as the cost signal.
- Set safety stock only where carrying cost is cheaper than expediting. High-runners with 55-day lead time and unpredictable demand justify buffer; low-volume items justify none and should be make-to-order or order-on-demand.
- Use coverage groups rather than item-by-item numbers: group items with similar lead time and variability, set minimum and maximum keys per group, and use a safety margin to add days to the 55-day lead time where the real risk is supplier delay rather than demand.
- Set the inactive 22% to zero minimum and zero maximum so planning stops recommending replenishment; handle remaining stock through write-down or reuse decisions, not through planning.
Measure expedite spend percentage and missed ship dates by item class, and review the coverage groups quarterly instead of leaving them for 14 months.
3. The planners want to keep firming planned orders by hand. What would you tell them, and which planning outputs would you have them work instead?
Read the worked answer
Firming by hand is a symptom of distrust, and it destroys the value of the plan: every firmed order becomes a constraint the engine cannot rebalance, so the next regeneration produces a worse plan, which increases distrust. Fix the plan and the firming habit shrinks.
Tell them:
- Firm only where a real commitment exists: a supplier promise, a customer commitment, or a production sequence already released. Not because the date looks wrong.
- Every firm action needs a reason recorded, so you can see whether the cause is a coverage setting, a lead time that is wrong in item master, or a genuine exception.
Work these instead:
- Action and delay messages from the plan, which show what changed since the last run.
- Pegging and planned order details to see why a quantity was proposed before overriding it.
- Coverage and forecast accuracy reports to correct the inputs that generate bad suggestions.
Report firmed orders as a percentage of planned orders weekly; a falling ratio is the evidence that the engine is being trusted and that Planning Optimization is delivering.