Simulation study, 400 SKUs, 1,825 days, three policies

Supply chain research poster example

A supply chain research poster has to show a trade off, because every policy that cuts stockouts holds more stock. This A0 landscape board simulates three reorder policies for 400 SKUs over five years of daily demand: policy B cuts stockout days by 44% for 12% more holding cost, and policy C costs nine times as much per extra day saved.

Create a poster with OneCraftA0 Landscape, printed at 1189 by 841 millimetres

The whole board

The poster at full size, exactly as it prints. Every number, citation and caption on it was written for this example, so the layout is being judged on real content.

Fewer stockouts, at what cost? Three reorder policies simulated for 400 SKUs over five years of daily demand

A. Lindqvist ¹, T. Mensah ¹

1 Operations Research Group, Corvell University

Abstract

A regional distributor reorders 400 stock keeping units with a fixed reorder point set from last year's average demand. We replayed five years of its daily demand, 1,825 days, through that policy and two alternatives. Policy A, the current rule, ran out of stock on 4.1% of SKU days. Policy B cut that to 2.3% for 12% more holding cost, and policy C to 2.0% for 29% more.

Simulation set up

400 SKUs, 1,825 days of demand, 3 policies, so each policy covers 730,000 SKU days. A: reorder point and fixed order size from the previous 12 month average. B: weekly forecast with safety stock set for a 97% service target. C: weekly review, order up to a 99% target. Lead times come from supplier records. Sources: Varga and Holt (2024) Reorder rules in practice, Operations Notes 18(2); Mendes (2025) Service targets and holding cost, Logistics Review 9(1).

HOW THE STUDY RAN

Demand history

1,825 days of orders for 400 SKUs, 2021 to 2025

Forecast

Weekly forecast per SKU, refitted every 4 weeks

Policy

Rules A, B and C set the reorder point and order size

Simulate

Replay each day with real supplier lead times

Score

Stockout share of SKU days and yearly holding cost

Figure 1. Stockout days and holding cost, indexed to policy A at 100. Bars left to right in each group: policies A, B and C.

Source: simulation, 400 SKUs, 1,825 days.

Results

Stockouts fell from 4.1% of SKU days under policy A, 29,930 days, to 2.3% under B, 16,790 days, and 2.0% under C, 14,600 days. Yearly holding cost rose from $312k under A to $348k under B and $401k under C. Moving from A to B removes 13,140 stockout days for $36k a year. Moving on from B to C removes a further 2,190 for another $53k, so each extra day avoided costs about nine times as much.

Discussion

Policy B is the sensible stop: most of the gain for little of the cost. The case for C rests on the 40 SKUs where a stockout loses a customer, and those can run at C while the rest run at B. Demand was historical, so a new product launch with no history is not tested here.

44% fewer

stockouts with policy B

2.3% of SKU days against 4.1%, for 12% more holding cost, $348k against $312k a year

Key finding

A forecast with a 97% service target removes most stockouts; pushing to 99% costs nine times as much per day saved.

Policy A, current rule: stockouts on 4.1% of SKU days

Policy B: 2.3% of SKU days, holding cost $348k a year

Policy C: 2.0% of SKU days, holding cost $401k a year

A to B saves 13,140 stockout days for $36k a year

Block by block

What each block on the board is for, in the order a reader walks it.

Title band and two authors: The title asks what fewer stockouts cost and names the 400 SKUs and five years of daily demand.
Abstract and simulation set up: The abstract gives the three stockout rates and cost rises. The set up section carries the method a protocol block would hold on other designs, plus two short citations.
How the study ran: The five step process chain with a detail line under each step. The chain holds five steps, so this method fits exactly.
Figure 1, indexed to policy A: Two groups, stockouts and holding cost, with bars for policies A, B and C in each, all indexed to A at 100.
Results and discussion: Stockout days per policy out of 730,000 SKU days, the yearly holding cost, the marginal cost of each step, and why B is the sensible stop with C kept for 40 critical SKUs.
Key number, key finding and points: The stat block holds 44% fewer stockouts with policy B. The callout states the trade off, and four checked points give each policy and the A to B saving.

What makes this board work

The process runs across the board

A five step chain spans the full width under the title: demand history, forecast, policy, simulate, score. A reader sees the whole method in one line before looking at a result.

Both costs on one scale

Stockouts and holding cost have different units, so Figure 1 indexes both to policy A at 100. Policy B reads 56 and 112, policy C 49 and 129, which makes the trade off a comparison of bar heights.

The marginal cost is spelled out

Moving from A to B removes 13,140 stockout days for $36k a year; moving from B to C removes 2,190 more for $53k. The results do that division so the reader does not have to.

Questions people ask

What should a supply chain research poster show?

The policies compared, the data they were tested on, both sides of the trade off and the cost of moving between options. Operations audiences want a decision, so show what each step buys: here 13,140 stockout days for $36k a year from A to B, against 2,190 for $53k from B to C.

How do I put two units on one chart?

Index them to a baseline. This board sets policy A to 100 for both stockouts and holding cost, so policy B reads 56 and 112 and policy C 49 and 129. Plotting 4.1% against $312k on one axis would flatten one series, and two separate charts would make the reader do the comparison.

Where does the protocol go on a layout without a methods block?

Into a section block, and the process chain carries the steps. This process flow design is one of two layouts without a protocol block, so the SKU count, days, policies and lead time source are in the simulation set up section, with each step summarised in the chain.

Can the process chain hold more than five steps?

No. The chain on this design draws exactly five, and a sixth is dropped without warning. If your method has more stages, merge two that always run together, as forecast and refit are merged here, or move the detail into the section text and keep the chain to the five a reader must see.

Why is the headline 44% rather than half?

Because 2.3% against 4.1% is a 44% cut, not a half, and the board should not round a result in its own favour. Stating the exact figure, with both rates in the detail line, lets a reader check it against the results in a few seconds.

Should a simulation poster include references?

Yes, at least for the policy rules and the service target method, since those are what a reader would look up. This design has no reference block, so two short author and year citations close the set up section. A layout with a reference block would hold three or four in full.

How do I count stockouts in a simulation?

Pick a unit and say it on the board. This study counts SKU days with no stock on hand: 400 SKUs over 1,825 days gives 730,000 per policy, so 4.1% is 29,930 days. Counting orders missed or customers lost would give different numbers, and readers in operations will ask which one you used.

Build your own in about a minute

The button below opens the generator with this use case already described. Change the wording to match your own, generate, then edit anything you like.

Make one like this

Sources

Written and checked by the OneCraft team. Last checked 4 October 2026.