Lead time variability: how to plan procurement when delivery times are uncertain

Lead time variability describes how much actual supply lead times differ between comparable orders. A supplier may quote ten days but alternate between deliveries that arrive within a week and others that take more than two. If procurement and production teams plan around that fixed lead time, they can face repeated changes, urgent orders and material shortages, even when the average remains ten days.

Measuring this variability helps you understand the uncertainty surrounding a delivery and decide how to protect against it through stock, earlier procurement or alternative sources of supply.

To use this information effectively, you need to know which dates are being compared, how much of the wait is attributable to the supplier and how much risk the business can accept for each item. This allows you to adjust planning to reflect actual supply performance.

What is lead time variability and why does it matter?

Procurement lead time measures the time from a defined starting point, such as placing an order, until the material is available. Its variability describes how much that time differs between transactions carried out under comparable conditions.

This should be distinguished from average lead time and adherence to promised delivery dates. A supplier may consistently deliver in twelve days while failing to meet a ten-day commitment: the lead time is stable, but the promise is unreliable. Another supplier may meet every commitment yet alternate between deliveries of eight, twelve or sixteen days because it confirms a different date for each order.

The average, variability and delivery performance answer different questions. The first helps estimate the usual duration; the second shows how widely lead times vary; the third assesses whether commitments are met. Informed procurement decisions require all three.

How to measure supplier lead time variability

Start with purchase order and goods receipt records. Before calculating any metrics, agree on a common definition and check that the dates reflect what actually happened.

Which dates and data should you record?

Record the date the order was placed, the supplier’s initial delivery commitment, any changes to it and the actual receipt date. If the material requires inspection or release, also keep the date on which it becomes available for production or sale. A component received on Monday but held for quality checks until Thursday cannot yet meet the requirements in the plan.

Decide how to handle partial deliveries. The first receipt can measure initial availability, but it does not represent the lead time for the full quantity. Use calendar days or working days consistently and check the calendars applied.

Separating order confirmation, preparation, transport and receipt helps identify where the wait arises. Monitor overdue open orders as well: analysing only completed orders can hide the longest delays. Track these separately until their final lead time is known.

What do the mean, standard deviation and percentiles tell you?

Average lead time summarises the duration of the transactions observed. To judge whether it is representative, you also need measures of dispersion and an understanding of the slowest deliveries.

Measure	What it shows	What it is used for
Mean	Average of the observed lead times	Estimating the usual duration
Standard deviation	Dispersion around the mean, in days	Comparing lead time stability
Coefficient of variation	Standard deviation divided by the mean	Comparing relative dispersion
Percentile	Lead time corresponding to a proportion of the sorted observations	Analysing the slowest part of supply

The coefficient of variation allows you to compare suppliers with different average lead times. Expressed as a percentage, it makes it easier to interpret dispersion relative to the usual duration. There is no universal threshold that separates acceptable suppliers from problematic ones.

P90 helps identify the lead time within which approximately 90% of historical deliveries arrived, depending on the calculation method. Percentiles can be estimated using different conventions, so use the same one when comparing periods. A historical percentile does not guarantee the same performance for future deliveries.

How to segment data by supplier, product and route

A supplier’s overall average may combine products available from stock with others made to order. It may also mix routes, plants of origin or quantities that require different lead times. Combining them reduces the usefulness of the metric.

Start with operationally comparable groups and break them down further when enough data is available. If there are very few receipts for a particular supplier and item combination, temporarily use a similar product family, make that limitation clear and review the estimate as new orders are added.

How variable lead times affect inventory and service

While waiting for a delivery, the business must meet demand from the material available. When the supply lead time changes, so does the quantity that may be consumed before the next receipt, even if sales remain stable.

Lead time variability and safety stock

Uncertainty in both demand and replenishment lead time contributes to stockout risk. If daily consumption of an item is ten units, each additional day of waiting means another ten units that must be covered by stock or a viable alternative.

Safety stock should reflect these uncertainties and the service target. It is useful to distinguish the cycle service level, which relates to the probability of completing a replenishment cycle without a stockout, from the percentage of units of demand fulfilled. These are different metrics, as explained in Peter L. King’s analysis of safety stock, published in APICS magazine.

When delays coincide with peaks in demand, treating them as independent events can underestimate risk. In these situations, test the policy against instances where both occurred together and against realistic scenarios.

The impact on reorder points and order dates

The reorder point depends on the demand expected during replenishment and the level of protection chosen. Keeping a ten-day lead time when recent deliveries have taken fourteen days may trigger procurement too late.

Decisions need to account for the inventory position: usable stock plus outstanding orders, less unfulfilled demand, with reservations accounted for without double counting. Arrival dates matter too; a delivery that arrives after the material is needed does not resolve the current shortage. Under periodic review, you must also allow for the time until the next opportunity to place an order.

Example: two suppliers with the same average lead time

A business analyses ten receipts of the same component from each supplier. All lead times are expressed in calendar days and measured up to the point when the material becomes available. The figures are hypothetical and illustrate the method.

Comparing actual delivery lead times

The observed lead times, sorted from lowest to highest, are as follows:

Metric	Supplier A	Supplier B
Observed lead times in days	9, 9, 10, 10, 10, 10, 10, 10, 11, 11	4, 6, 7, 8, 9, 10, 12, 14, 14, 16
Average lead time	10 days	10 days
Sample standard deviation	0.67 days	3.92 days
Coefficient of variation	6.7%	39.2%
P90 using the nearest-rank method	11 days	14 days

To calculate P90, multiply the number of data points by 90%, round the result up to a whole number and take the observation at that position: here, the ninth of ten. Supplier A’s deliveries fall between nine and eleven days; supplier B has a much wider range.

Both have the same average, but the protection required needs to be assessed differently. With only ten observations, these results describe the example and are insufficient to establish a definitive policy.

What changes in procurement planning?

Assume constant consumption of ten units a day, continuous review and no overlapping orders. To isolate the effect of lead time, we exclude demand variability and order quantity constraints. The business tests an illustrative rule that covers consumption up to the observed P90.

Result of the rule	Supplier A	Supplier B
Consumption during the average lead time	100 units	100 units
Consumption up to P90	110 units	140 units
Additional protection above the average	10 units	40 units
Illustrative reorder point	110 units	140 units

Under this rule, B requires thirty more units of protection than A. That difference helps assess its price, supply terms and stock holding costs. It may also justify negotiating more consistent deliveries or securing an alternative source.

Using P90 does not guarantee a future service level of 90%. The rule must be tested against more data, the way replenishment works in practice and its results. If demand also varies, its distribution during the lead time will need to be analysed.

How to plan procurement with variable supply lead times

Once you have measured variability, you need to decide how to incorporate it into the plan. The answer will depend on the service commitment, the cost of a shortage and the options available for each material.

Which lead time should you use as a planning parameter?

The recent average can be a reasonable basis for estimating arrivals when performance is stable. If some deliveries take much longer than usual, examine the distribution and check whether that estimate leaves too many requirements unmet.

The period analysed should reflect future conditions. A change of route or an increase in capacity may make some historical data outdated. Keep the earlier information to understand the change, but avoid automatically combining it with the new pattern.

Distinguish the expected receipt date from the buffer used to protect service. If both are captured in a single lead time, it becomes difficult to tell whether an order is being placed earlier because of an expected delay or a decision to provide extra cover.

How to adjust replenishment without duplicating protection

Extending the lead time in the system, increasing safety stock and manually bringing orders forward can duplicate protection against the same risk. Before changing parameters, identify which uncertainty each adjustment covers and check their combined effect.

Also review the ordering calendar, minimum order quantities and outstanding deliveries. Ordering earlier may be necessary, but it commits cash and can increase exposure to changes in demand. Assess the policy both by the service it achieves and by the stock levels and urgent orders it generates.

How to reduce variability with suppliers

Breaking lead times down allows you to address specific causes. If the wait arises during confirmation, agree response times. If it stems from transport, changing frequencies or routes may help. If material is held at goods receipt, that internal process needs improvement.

Sharing future requirements and reviewing available capacity can help suppliers make more realistic commitments. Where dependence remains high, supplier risk assessment helps evaluate alternatives, their qualification and the volumes they could handle. Check their viability before relying on them in the plan.

When to review lead times and update parameters

Review lead times at a frequency that reflects how often you order and how critical the supply is. Alongside this regular schedule, certain warning signs warrant an assessment before the next planned review.

What signs suggest a change in supplier performance?

A sustained increase in dispersion, a shift in the average or several deliveries outside the usual range may indicate a change. Also check whether order quantities, routes or the product mix have changed.

An exceptional incident deserves follow-up, although it may not justify changing every parameter. Investigate its cause and the likelihood of recurrence. A supplier scorecard can bring these signals together with delivery performance and quality to help prioritise action.

Who should approve and check adjustments?

Procurement should validate the supplier’s terms and commitments. Planning should translate them into parameters and check their consequences; Finance should be involved when a change requires more capital than has been authorised. Each adjustment needs an owner and a review date.

Record the previous value, the new value, the reason for the change and the items affected. Then check whether service improved and urgent orders decreased, and review what happened to stock levels. This follow-up allows you to retain, adjust or remove the additional protection.

Better procurement planning starts with understanding actual lead times

Lead time variability explains why an apparently reasonable average can still lead to material shortages. Measuring it with comparable data allows you to adjust replenishment and assess which risks are best covered by stock and which require improvements to supply.

When you manage many items, supply chain planning software makes it easier to keep dates, outstanding orders and requirements aligned. Connecting this information helps you review the consequences of a decision before passing it on to procurement or production.

SCP Studio connects demand, procurement and production so that these decisions form part of a shared plan. To explore how this could apply to your materials and suppliers, you can request a demo with our experts.