Spare part safety stock calculator
For a slowly consumed spare part, see the expected usage over the lead time, the smallest stock that meets the target service level, and the holding and stock-out cost of that stock.
Cost comparison (optional)
The calculation runs in your browser; the values you enter are not sent anywhere.
- Stock needed (reorder point)
- 2 units
- Expected usage over the lead time μ
- 0.493
- Safety stock (stock − μ)
- 1.51
- Probability of a stock-out during the lead time
- 1.387%
- Average wait for a stock-out
- 24 years
Holding cost versus downtime cost
| Target service level | Lowest total cost | |
|---|---|---|
| Stock (units) | 2 | 3 |
| Service level | 98.613% | 99.833% |
| Expected stock-outs per year | 0.0416 | 0.005 |
| Annual holding cost | €722 | €1,082 |
| Annual expected stock-out cost | €1,498 | €180 |
| Total annual cost | €2,221 | €1,262 |
The stock with the lowest total cost can differ from the target service level: if downtime is cheap less stock is economic, if it is expensive more. For critical parts, safety and legal requirements can also decide beyond the economic calculation.
Sensitivity: stock and cost by service level
| Service level | Stock | Safety stock | Total |
|---|---|---|---|
| 90 % | 1 | 0.51 | €9,881 |
| 95 % | 2 | 1.51 | €2,221 |
| 99 % | 3 | 2.51 | €1,262 (lowest total) |
| 99.5 % | 3 | 2.51 | €1,262 (lowest total) |
| 99.9 % | 4 | 3.51 | €1,460 |
■ lowest total
Assumptions
- Demand is Poisson distributed: usages arrive independently at a constant average rate. The assumption weakens for equipment that consumes several parts at once or where parts are replaced in batches.
- The lead time is fixed; supplier delay is not in the model. With uncertain supply, enter a value towards the long end.
- The stock-out probability per order cycle is taken as P(D > s); demand is backordered during a stock-out.
- Average stock on hand is approximated as safety stock + order quantity / 2 (at least 0).
μ = annual usage × lead time / 365 · P(D ≤ s) = e^(−μ) Σₖ₌₀ˢ μᵏ / k! ≥ service level · safety stock = s − μ · stock-outs/year = (annual usage / Q) · P(D > s)
The result is a model estimate; actual demand, supply delays and part life can differ. With slow usage, little history makes the estimate uncertain. For critical parts the decision should be taken together with manufacturer recommendations and a risk assessment.
Let's look at which parts on your spares list warn before they fail, and what that does to your stocking policy.
Request a conversation01
How to use it
A
Enter annual usage, lead time and target service level (choose the criticality for a starting suggestion).
B
Read the stock needed, the safety stock and the stock-out probability.
C
Enter unit price, holding rate and downtime cost to compare holding and downtime cost; the sensitivity table shows what each service level costs.
02
Why Poisson for slowly consumed parts?
Spare parts such as bearings, boards and valves are used a few times a year. For such rare, random events the number of usages is generally close to a Poisson distribution. Its single parameter is the expected usage over the lead time μ: annual usage × lead time / 365.
The classic safety stock formula based on the normal distribution (z × σ × √L) works for large, smooth demand; when demand is 1–2 units a year, a continuous distribution gives fractional and misleading results. The Poisson model gives a correct, discrete answer in units.
03
Service level and the stock needed
The service level is the probability that demand during the lead time is met from stock on hand. The tool finds the smallest whole number s with P(D ≤ s) ≥ target. This s is the total stock on hand and on order at the moment an order is placed (the reorder point). Safety stock is that value minus the expected usage μ.
Raising the service level is increasingly expensive: going from 95% to 99.9% often means only one or two extra units, which on a high-priced part is a substantial amount of tied-up capital. The sensitivity table shows the trade-off.
04
Holding cost versus downtime cost
Holding stock costs capital, storage and obsolescence; running out costs downtime and urgent procurement. The tool adds the two and also shows the stock that minimises the total. It need not coincide with the target service level.
A monitoring system that warns before failure changes this equation: if a failure is noticed a week ahead the part is procured on a planned rather than urgent basis, the downtime of a stock-out shortens and a lower stock may be enough.
FAQ
- What is safety stock?
- Additional stock held above the expected usage over the lead time, against demand coming in higher than average. Here safety stock = stock needed − expected usage (μ).
- What does a 95% service level mean?
- That the probability of the part being in stock during the lead time is at least 95%; in other words you accept running out in about five of every hundred order cycles. For a part ordered only rarely this can mean a stock-out once in decades.
- Should I hold stock for a part used once a year?
- If the part is critical and the lead time long, often yes; the decision depends on the balance between the part's price and the downtime cost. The tool's cost comparison shows that balance. For a cheap part that causes downtime, one unit in stock is usually economic.
- How do I enter annual usage when I have several identical machines?
- Enter the total annual usage of all equipment that uses the same part; shared stock needs fewer units than separate stock for each machine.
- What if the lead time is uncertain?
- The model assumes a fixed time. To allow for uncertainty enter the time towards the long end (for example close to the longest of your recent orders) and try a few values for sensitivity.
Build your spares policy on failure data
Let's look at which parts warn before they fail, and what that means for spare stock and downtime.