Release gates, exceptions, and change control

A proposal is permission to investigate; a released order is permission to execute under the organization’s rules. Northstar uses a release gate to avoid sending attractive but impossible quantities to the shop. The gate checks the item version, material, usable capacity, due date, and accountable owner. The shared baseline needs twenty nine additional good bottles. Supplier availability is twenty four good-unit equivalents, manufacturing can support thirty two, and shipping can handle thirty. These figures are expressed on a comparable good-unit basis for this screen. The smallest limit is twenty four, leaving a five-unit gap before recovery. Do not confuse that screen with the later start calculation. Once enough material for the target is confirmed, twenty nine expected good units still require thirty starts at ninety seven percent yield. A supplier limit stated in raw bodies would need conversion before comparison with good units. The exercise deliberately labels the units to avoid that trap. A release checklist should have evidence, not only check marks. Record the receipt confirmation, approved BOM version, resource slot, and shipping cutoff. An owner should know which evidence becomes stale when demand changes. A changed due date can invalidate the slot even if the quantities remain unchanged. Use a time fence to describe how changes are handled, not to pretend that changes cannot happen. Northstar’s next two days are frozen: any quantity or date change requires planner and shop approval. Outside the fence, the planner may revise proposals, but must still rerun dependent material and capacity checks. When the gate fails, record a specific exception. State the missing quantity, affected demand, latest useful decision time, and candidate remedy. The system should preserve the original need even if it releases only the feasible part. Otherwise a partial release can make an unresolved customer obligation disappear from view. A planning exception marks a mismatch between the intended plan and current evidence. It can be a late receipt, a yield loss, an overload, or a version change. Start by identifying which assumption changed. Rebuilding every number without preserving the reason makes it hard to tell whether the revision solved the actual problem. Northstar plans two shipments. There are ten usable bottles at the start of week one, a twenty-bottle receipt in week one, and a twenty-five-bottle receipt in week two. Demand is twenty eight and twenty four. The baseline ends week one with two and week two with three. Both due dates are covered. If ten bottles from the first receipt slip to week two, week one has only twenty against demand of twenty eight. Eight units become backlog. In week two, the delayed ten and the planned twenty five arrive. After serving eight backlog and twenty four new demand, three remain. The final balance is unchanged, but service in week one is worse. A total-stock report may therefore miss the most important event. Report on-time units separately from eventual fulfillment. A promise about week one cannot be repaired retroactively by an abundant week-two balance. A planner should give customer-facing teams the affected quantity and credible revised timing. Now pause and test the plan with your own hands. Draw a row for each week. Name the unit at the top of each row. Put the input in one place and the result in a new place. Keep a copy of the first plan. Change just one fact, then run the sums once more. Ask what moved, why it moved, and who must act. If a date is too late, say so. If a part is not there, show the gap. If a count is based on an average, do not call it a sure thing. A plan is useful when the next person can check it and act on it. Write a short note that states the need, the choice, the risk, and the next check. Read it out loud to a peer. Ask that peer to find one weak claim. Use the feedback to make the plan clear. Each number should have a unit, a time, and a source in the case. Each change should have a reason. Each open gap should have an owner. These habits help a small team see the same facts and make a sound choice.

1. The planning question

   Original card with the two lesson goals.

   • Turn proposals into executable orders with a small release gate.
   • Replan with an audit trail and explain customer consequences.
   • Northstar fictional case
2. Inputs and units

   Original table highlighting input quantities and dates.

   • Firming
   • Time fence
   • State timing
3. First calculation

   Original equation card showing the first worked case.

   • Base feasible[24,30,18,28];gaps[5,5,2,3].
   • Supplier-only addition yields[29,30,20,28];gaps[0,5,0,3].
   • Shipping-only addition yields[24,30,18,31];gaps[5,5,2,0].
4. Second calculation

   Original table showing the second worked case.

   • Disruption on-time units20+24=44 of52,84.62%;8 prior units fulfilled late in week2.
   • Expedite4:week1 on-time24,backlog4;week2 ending3;on-time48/52=92.31%.
   • Full expedite and week2 demand30 gives3 backlog at end week2 and0 stock;on-time49/58=84.48%.
5. Stress the plan

   Original side-by-side baseline and revised result cards.

   • Change one input
   • Recompute dependent rows
   • Expose the gap
6. Commit with evidence

   Original checklist with an owner and next review date.

   • Retain assumptions
   • Check the result
   • Name the next action