How Supplier Lead Times Affect Your Inventory Forecast

Most reorder-point math treats lead time as a fixed number you plug in once. Ask a merchant what their supplier's lead time is and they'll usually give you the number from the last email — not the number their last eight orders actually took.

Why lead time breaks forecasts

The reorder point formula and the target stock level formula both use lead time the same way: average daily sales × lead time is the demand you need covered before the next shipment lands. Get the lead time wrong and everything built on top of it is wrong by the same margin, even if your sales forecast was perfect.

The failure is quiet. A supplier that used to take 10 days and now takes 18 doesn't send a notice — you find out when the safety stock that was supposed to be a buffer becomes the only stock left, days before the shipment actually arrives.

Measuring real lead time

Start from your own purchase order history, not the supplier's quoted number. For each of the last 8–10 orders on a SKU, record the date you placed the order and the date it was fully received and available to sell. Average those gaps — that's your real lead time, and for most suppliers it's longer than the quoted figure, because the quote usually covers production only, not shipping and receiving.

Variability, not just length

The average is half the picture. Look at the spread across those same 8–10 orders. A supplier that runs 10, 11, 12, 13, 11 days is a very different planning problem than one that runs 6, 9, 14, 18, 8 — even if both average out to roughly 11 days. The first is a length problem you plan around once. The second is a variability problem, and it's the one that actually causes stockouts, because you can't know in advance which order will be the slow one.

Statisticians call that spread the standard deviation of lead time, written σ_LT. It's an exact number you can calculate from your own order history, not a feeling about whether a supplier is "reliable."

Scaling the buffer

The safety-stock buffer attributable to lead-time variability alone has a real statistical formula, not a rule of thumb:

Safety stock (lead time) = Z × σ_LT × Average daily demand

Z is the service-level factor, and this is the single most common place a safety-stock calculation goes wrong: Z = 1.65 for a 95% service level, not 1.96. Safety stock protects against running out on one side only — a one-tailed statistic — and 1.96 is the two-tailed value most generic statistics content hands you. Using 1.96 here overstates the buffer.

Applied to Cedar & Fig, 250g: average daily sales of 5 units, and suppose the last 8 orders' lead times worked out to a standard deviation of 2 days.

Safety stock = 1.65 × 2 × 5 = 16.5, rounded to 17 units. That's not the same 30-unit buffer used elsewhere on this site for the same SKU — the 30 came from a simpler days-of-cover rule of thumb, a legitimate way to set a buffer without knowing σ. The 17 is how much of that buffer lead-time variability alone would statistically justify at a 95% service level. The full combined formula, folding in demand variability too, is one square root away — the same math the reorder point guide's safety-stock section already covers.

The relationship isn't linear, either. Push the service level from 90% (Z = 1.28, ≈13 units here) to 99% (Z = 2.33, ≈23 units), and the buffer grows by roughly three-quarters for just 9 more percentage points of service level. A 100% service level isn't on that table because it's statistically unreachable — there's always some chance of a longer-than-usual delay, no matter how much stock you hold.

Splitting orders across suppliers

Buying the same SKU from two suppliers spreads out lead-time risk. A delay from one no longer means the whole order is late, because you're no longer fully exposed to a single supplier's variability. It isn't free, though: two minimum order quantities and two relationships to manage, and it does nothing if both suppliers ship through the same port during the same disruption. Nobody has published a defensible number for how much risk splitting actually removes: treat it as a tool for your highest-risk SKUs, not a formula.

Tracking σ_LT by hand means pulling every PO's placed and received dates for every supplier, every quarter, and recalculating. Most stores do this once, if ever, and then keep using a stale lead time for a year. StockCue uses the lead time you set for a supplier as an input to every forecast on every plan, including Free. On the Scale plan, it computes each supplier's actual average lead time from your own purchase-order and receiving history automatically, so the number feeding the formula is measured, not remembered.

Frequently Asked Questions

How does supplier lead time affect an inventory forecast?

Lead time is one of the three inputs to both the reorder point and the target stock level — it sets how much demand you need to cover before new stock arrives. A forecast that uses last quarter's lead time while the supplier is now running slower will understate how much buffer you actually need, and the store stocks out before the shipment lands.

What's the difference between a long lead time and an unreliable lead time?

A long lead time is a length: 30 days instead of 10, and you can plan around it once you know the number. An unreliable lead time is a spread: sometimes 8 days, sometimes 22, and no single number describes it well. Two suppliers can share the same average lead time and carry very different risk if one is consistent and the other isn't.

Should safety stock be based on demand variability or lead time variability?

Both feed into the same combined safety-stock formula, and neither dominates as a rule; it depends on the SKU. In the one detailed worked example available from safety-stock literature, demand variability was the larger factor, but lead-time variability's contribution scales with sales volume, so a high-volume SKU with an unreliable supplier can have lead time as the bigger driver. Calculate both rather than assuming one.

Does ordering from two suppliers reduce lead-time risk?

It can, because you're no longer fully exposed to one supplier's variability, but it isn't free: dual sourcing usually means two minimum order quantities and two relationships to manage, and it doesn't help if both suppliers share the same shipping chokepoint. No sourced figure exists for how much risk splitting removes; treat it as a mechanism, not a guaranteed reduction.

For the reorder point formula this post assumes, see the reorder point formula guide. For sizing the total quantity you need, see how much inventory you should keep, or the complete guide to Shopify inventory forecasting for the full process.