Closed-loop supply chain planning is a planning cycle that learns from its own results. The system recommends decisions, the business executes them, and the gap between what was planned and what happened is measured and fed back into the next plan. In sustainability writing, “closed-loop supply chain” usually means returns, repair and remanufacturing, a separate subject from the planning feedback loop covered here.
What Does Closed-Loop Supply Chain Planning Mean?
Every planning cycle has the same four steps. In an open loop, the plan goes out and the next cycle starts from the same assumptions, whatever happened in between. In a closed loop, the last step changes those assumptions.
- Plan: the system produces recommendations, such as purchase orders, store replenishment quantities, allocations, production runs and truck loads.
- Execute: approved decisions go to the ERP, WMS or carrier.
- Measure the gap: compare executed decisions with recommended ones, and actual results with expected ones.
- Feed it back: correct the inputs, parameters and rules that caused the gap, then plan again.
The loop runs at the speed of the decision. Store replenishment can close it every day. A production plan closes it every week, and a network design every year or so.
Measuring the Gap Between Plan and Execution
Measure two gaps separately, because they point to different fixes.
The decision gap shows whether planners executed what the system recommended. The usual measure is user acceptance rate: the share of executed planning decisions that go live with no planner edits. When a planner changes a quantity, moves a date or rejects an order, that decision counts as edited. Solvoyo explains the measure and how it is calculated on its user acceptance rate page.
The outcome gap shows whether the executed decision produced the result the plan expected. Typical measures are forecast accuracy and bias, plan adherence, fill rate or OTIF, stock cover and cost per unit shipped.
| Gap | Question it answers | Typical measures | What gets fed back |
|---|---|---|---|
| Decision gap | Did planners run the recommendation as it was? | User acceptance rate, edit reason codes | Missing constraints, business rules, data errors |
| Outcome gap | Did the executed decision deliver? | Forecast error and bias, plan adherence, fill rate, OTIF, cost per unit | Forecast models, lead times, safety stock, service targets |
Each edit needs a reason attached. An edit with no reason gives the team nothing to act on. An edit tagged “supplier short on pallets” or “promotion not loaded” tells the team what the model is missing.
Feeding the Gap Back Into the Plan
Feedback works when it changes something the next plan uses. In practice the changes fall into three groups.
- Parameters: lead times relearned from actual receipts, safety stock retuned when sales or supplier OTIF variance rises, and forecasts corrected for periods when stock-outs suppressed demand.
- Rules and constraints: a pack size, shelf capacity or truck limit that planners kept fixing by hand gets added to the model, so the fix stops repeating.
- Data: master data errors flagged by diagnostics are corrected before the next run, instead of being patched in each plan.
Once a parameter or forecast changes, the plan should re-run on its own. If someone has to remember to restart it, the loop stays open.
Which Planning Decisions Should You Automate First?
Closed-loop planning makes automation safer, because every automated decision is measured. Good first candidates share four traits:
- High volume: thousands of similar decisions per cycle, where planner time is spread thinnest.
- Repeatable: the same decision with the same inputs each day or week, such as store replenishment or DC purchase orders.
- Measurable: the result shows up within a cycle or two, so the loop can close quickly.
- Bounded cost of error: one wrong order line costs little and can be corrected next cycle.
Store replenishment, purchase orders, allocation, order fulfillment and load building usually fit. Solvoyo’s published acceptance rates include 81% for Studenac’s store replenishment in its first year and 93% at a mid-market CPG company, both listed on Solvoyo’s user acceptance rate page.
Some decisions should stay with planners, with the system computing options for them:
- Decisions that depend on information outside the data, such as a supplier negotiation or a one-off commitment to a key customer.
- Low-frequency, high-stakes choices, such as opening a DC or changing a network, where people choose between scenarios the model has costed.
- Policy calls, such as which service level a category should target or how much risk to accept on a launch.
Exception Management: Where Planners Spend Their Time
Exception management decides which recommendations go straight to execution and which wait for a planner. Rules and tolerances set the line. A recommendation inside its tolerance runs. One that breaches a threshold, such as a forecast deviation, a capacity limit, a service risk or a cost impact above a set amount, goes to a planner’s queue.
Three habits keep the queue useful:
- Rank exceptions by business impact, so the largest service or cost risks are handled first.
- Close each exception with a reason, which becomes part of the feedback.
- Review the thresholds. If planners approve the same type of exception unchanged week after week, its tolerance is too tight and it can run automatically.
This is how the share of automated decisions grows: one workflow at a time, widened as acceptance holds. A supply chain control tower is the usual place where planners see and work these exceptions. Planners also need to see why the system recommended what it did, which is the subject of our post on explainable AI in supply chain planning.
How Solvoyo Approaches Closed-Loop Planning
Solvoyo is a supply chain decision automation platform built on optimization and machine learning. One solve weighs cost, service, capacity and lead time together, and constraints found through planner edits are added to that model. That recommendation quality is why over 95% of plans execute untouched, sustained across enterprise clients.
Solvoyo tracks acceptance on every decision with a reason code for each edit, next to the business KPIs those decisions should move. Clients choose which decisions run on their own, by workflow, location and decision type. Customer Success engineers review the most edited SKUs and add the missing cause to the model in 3 to 6 week sprints. The supply chain analytics and visibility layer shows exceptions in one control tower, traces each to its root cause and ranks it by business impact, on the same data model the plans use. To see the loop on your own data, request a demo.
Frequently Asked Questions
Closed-loop supply chain planning is a planning cycle in which the system recommends decisions, the business executes them, and the gap between planned and actual results is measured and fed back into the next plan. The feedback changes forecasts, parameters such as lead times and safety stock, and the rules and constraints in the planning model.
Start with decisions that are high volume, repeatable, quick to measure and cheap to correct, such as store replenishment, purchase orders, allocation and load building. Keep low-frequency, high-stakes choices and decisions that depend on information outside the data with planners, and widen automation as acceptance holds.
No. A closed-loop supply chain usually refers to product returns, repair, remanufacturing and recycling flowing back into the forward supply chain. Closed-loop planning refers to the feedback loop between a plan and its execution, where measured results correct the next plan.





