Case study
Company X: The Weekend Reorder Decision
A long-form walkthrough of one inventory call before a city festival at a multi-location retailer: three teams, one hero SKU, and a reorder point that looked fine until marketing turned on paid traffic.
The company
Company X runs twelve retail locations in one metro area. Most revenue comes from drinks, but the hero product is a house espresso blend sold by the bag for home brewing. It is high margin, high visibility, and the SKU marketing promotes during events.
Inventory is centralized: one warehouse feeds all stores. When the warehouse runs low, stores post “sold out” signs within days. Last year, during a spring street festival, the blend stocked out on Saturday afternoon. Paid social was still running. Conversion collapsed, store managers got angry, and finance still had to pay for ads that pointed at empty shelves.
This year the festival is back. The general manager asked for a single number before Monday's planning meeting: what reorder point protects the weekend without trapping cash in the warehouse?
Monday morning: three views
Operations wanted to keep the current reorder point at 900 bags. Normal weeks had been fine. Their argument: we already carry more than a week of average demand; raising the number freezes working capital.
Finance pushed the opposite. Cash was tight after a equipment upgrade. They proposed lowering the reorder point to 780 bags to rotate inventory faster. Their spreadsheet showed lower carrying cost per month.
Marketing had booked festival ads expecting a 35 to 40 percent traffic bump on the blend landing page. They asked for 1,200 bags at reorder, citing last year's stockout. “We are not running ads into a sold-out page again.”
Everyone had a reasonable story. Nobody had written down the probability of stocking out during supplier lead time under each policy.
The numbers on hand
The analyst pulled twelve weeks of outbound data for the blend. Non-festival weeks averaged about 118 bags per day across the chain. Festival-adjacent weeks ran closer to 145. For planning, the team used 132 bags per day as a central estimate with higher volatility than usual: standard deviation about 34 bags per day, reflecting weekend spikes and uneven store sell-through.
The roaster-supplier quote was clear: seven calendar days from PO to dock, assuming no transport delay. That lead time is the window where stock has to cover uncertain demand once the reorder fires.
Plain-language lead time
Lead time is not how long until the truck arrives on a good day. It is how long you must survive on warehouse stock after you hit the reorder button. During those seven days, demand can swing above or below average. The reorder point is your buffer against that swing.
Work the case: test each policy
Before debating personalities, the team modeled four policies on the table. Use the explorer below the way they did in the meeting: pick a preset or drag the slider, and read stockout probability and service level side by side.
Work the case: pick a reorder policy for Company X
The supplier needs seven days to deliver. Demand during that window is uncertain, especially during the festival. Move the reorder point and read stockout probability before the Monday decision.
Ops (current policy): Worked on normal weeks last quarter.
Mean lead-time demand
924 bags
Reorder point
900 bags
Stockout probability
60.5%
Service level
39.5%
Stockout risk vs reorder level
At 900 bags, stockout risk is 60.5% during the 7-day lead time. Expected lost margin from shortfalls is about $138 per reorder cycle — before counting ad waste and customer complaints.
Policy comparison table
The table summarizes the four proposals. Stockout risk is the probability that demand during the seven-day lead time exceeds the reorder point. Service level is one minus that risk: the share of reorder cycles where you expect to cover demand through the window.
| Policy option | Reorder point | Stockout risk | Service level | Safety stock | Read |
|---|---|---|---|---|---|
| Status quo | 900 | 60.5% | 39.5% | -24 bags | Too exposed for a hero SKU during a promo week. |
| Finance cut | 780 | 94.5% | 5.5% | -144 bags | Too exposed for a hero SKU during a promo week. |
| Festival buffer | 1,200 | 0.1% | 99.9% | 276 bags | Strong service level; watch working capital. |
| High service | 1,350 | 0.0% | 100.0% | 426 bags | Strong service level; watch working capital. |
What the math was saying
Mean demand over seven days is roughly 924 bags (132 × 7). A reorder point of 900 sits slightly below that mean. On a typical week you might scrape by. On a volatile festival week, sitting below the mean is a deliberate bet that variance will stay kind — a bet last year already lost once.
Central estimate
For Company X: 132 × 7 ≈ 924 bags expected before the next shipment lands, before any safety buffer.
Risk definition
Higher reorder points push stockout probability down. Lower points free cash but expose you to empty shelves when daily demand lands in the upper tail.
Operational readout
Marketing cared about service level on the hero SKU because ads and email were already committed. Finance cared about the cash tied up in bags sitting in the warehouse.
The full formula write-up lives in the stockout probability post. This case study is the narrative layer on top: names, stakes, and a meeting where the number had to be defensible.
The hidden cost of stocking out
Finance's spreadsheet only counted warehouse carrying cost. It did not count lost margin on the blend when stores stock out during paid traffic, or the support tickets when customers arrive after seeing an ad. A rough margin of $9.50 per bag makes even a small expected shortfall expensive during a promo week.
That is where expected value thinking helps without over-modeling: compare the cost of holding extra bags for seven days against the expected loss from stocking out. You do not need a perfect model. You need both sides of the trade on the same slide.
See expected value for comparing bets and percentiles for reading demand tails if you want the foundations behind these reads.
What they decided
The GM rejected the finance cut (780 bags): stockout risk landed near one in four, too high for a promoted SKU. Status quo (900) was still double-digit risk once festival variance was in the model — better than 780, but not good enough for a repeat of last year.
They set reorder at 1,200 bags for the three weeks around the festival, then revert to 950 for normal weeks after post-mortem data came in. That put stockout risk under ten percent for the promo window while keeping a written plan to unwind extra safety stock once ads turned off.
Marketing got protection for the campaign. Finance got a time-bound policy, not a permanent step-up in average inventory. Ops got a number they could put on the PO instead of “we will watch it daily.”
What they watched the next week
Daily sell-through by store (not chain average alone). Two slow locations can hide a hot store burning through allocation. They also tracked landing-page conversion on the blend SKU so marketing could pause spend within hours if inventory signals slipped, instead of after weekend reviews.
After the festival, actual demand during lead time was compared to forecast. That feedback loop updates the mean and standard deviation for the next event — the parameters in the explorer are not permanent truth.
Habits worth keeping
Put reorder points in planning docs with explicit stockout probability, not only bag counts. Separate festival policies from baseline policies with start and end dates. Pair inventory calls with whoever is spending to drive demand — ops and marketing should see the same risk read.
Case studies like this are not about finding the one correct reorder point for every business. They are about making the trade visible before the truck is late and the ads are already live.