Safety stock without lead-time data: three fallback methods

The standard safety stock formula refuses to run without two inputs most small catalogs don't have: average lead time and demand variability. A supplier with no shared history, PO dates buried in email threads, a SKU too new to have a track record — and the reorder decision doesn't wait for clean data. You still need a buffer number today. The three fallbacks below each run on a different subset of what you have — a days-of-cover proxy when you know only your sales rate, a quote-based estimate when the supplier gave you a number, and guess-and-check when nothing is trustworthy yet. Each has a fixed upgrade path the moment real data arrives. The full formula they approximate is covered in the safety stock guide; this article handles the no-data case.

First, check what you actually have

Most "no data" situations are narrower than they feel. Triage the situation before picking a method:

The order and receive dates you'd need for the first row live on the purchase order record (Shopify Help Center, accessed November 2026). The middle row is the most common: you know what the supplier promises and nothing about what it delivers. That's Fallback 2's home turf, and the quote's known biases are manageable with padding.

Fallback 1: the days-of-cover proxy

The crudest method, and the only one that needs no supplier information at all: express the buffer as days of average sales rather than units. Buffer = N × daily demand. Choose N by consequence — roughly 10–14 days for products that must not run bare, 5–7 for steady mid-tier items, zero for the tail (illustrative starting ranges; set them from your own tolerance for bare shelves).

Worked: PIN-EN-048, an enamel pin, sells 5.2/day on your trailing average (illustrative). A 12-day buffer = 5.2 × 12 = 62.4 → 63 units. That buffer sits under the reorder line; and if you don't know the lead time at all, the days-of-cover reading doubles as the trigger itself — reorder whenever on-hand drops below N days of cover, which is the arithmetic in the days-of-cover guide.

Strengths: computable in a minute, self-explanatory to a team, and impossible to get wrong in units. Weakness: it's blind to replenishment reality — 12 days of buffer is comfortable behind a 10-day supplier span and useless behind a 40-day one. Use it while you lack the span, not instead of learning the span.

Fallback 2: the quote-based estimate

When the supplier gave you a number, spend it — with padding, because quotes habitually cover production alone. The pattern: planning span = padded quote; buffer = planning span × daily demand × risk share. The risk share stands in for the variability data you don't have: 50% is a serviceable illustrative rule for a first cycle (hold insurance equal to half the demand you expect during the wait), 25% for suppliers with a domestic track record, 75% when the route is long and unproven.

Worked: SEED-TOM-14, a tomato seed packet, sells 2.6/day. The supplier quotes 30 days including freight to your door — confirm that scope first, because a quote that stops at the port changes everything (the decomposition questions for a first PO are in the new-supplier lead-time guide). Pad 30 by 25% → 38-day planning span. Expected demand during the wait: 2.6 × 38 = 98.8 ≈ 99. Buffer at a 50% risk share = 50 units. Reorder line = 99 + 50 = 149 units. Native purchase orders in the admin carry an expected-arrival field, so the padded date can live in the same record as the order (Shopify Help Center, accessed November 2026).

Real limits: this method inherits every flaw of the quote plus your chosen padding. It's the best estimate available, and it's also the one carrying the most assumptions — which is exactly why the correction loop in Fallback 3 matters even here.

Fallback 3: guess-and-check

When nothing can be trusted, pick a defensible starting buffer and let the market correct it. Start at two weeks of demand. Watch the SKU weekly in a saved view — Shopify has no native low-stock email alert, so threshold monitoring means saved views plus Shopify Flow if you automate (Shopify Help Center, accessed November 2026) — and adjust after every bare spell or near miss.

Worked: LIGHT-BK-02, a bike light, sells 1.9/day (illustrative). Starting buffer = 2 weeks = 1.9 × 14 = 27 units. Two bare cycles inside a month → raise to 40. Six quiet weeks at 40 → hold, and test 35 next quarter if cash tightens. Each adjustment is one cell in a spreadsheet, and the loop converges on the true number within a season or two. Shopify's native counts carry a planned date; the weekly verification cadence is a calendar entry you own.

The prerequisite is accurate counts: buffers calibrated against phantom stock correct toward the wrong number, so keep the count discipline from the Shopify counting guide underneath this method. The other cost is tuition — you learn by going bare a few times, which is cheap on the tail and expensive on a bestseller.

Which fallback, and when to stop

Run the cheapest method your data supports: proxy when you know only the sales rate, quote-based when a quote exists, guess-and-check when nothing does — and expect to run two methods side by side during a supplier's first quarter, since the quote arrives before the history does. The stopping rule is the same for all three: the moment a supplier has six or more POs with recorded order and receive dates, retire the fallback and compute the real formula. Measured lead time and measured variability beat every proxy, and extracting that history from a PO record is a spreadsheet exercise, not a project. Keep the fallback numbers afterward as a sanity band — if the formula ever suggests a buffer far below what the fallback survived on, check the inputs before trusting it.

Common questions

How much safety stock should I keep with no lead-time data? Use the days-of-cover proxy: 10–14 days of average daily sales for products that must not run bare, 5–7 for steady mid-tier, none on the tail. Once the supplier gives you a quote, switch to the padded-quote method, and retire both the moment you have six or more measured POs.

What is a days-of-cover buffer? A buffer expressed as days of sales instead of units: multiply your average daily demand by a chosen number of days. A 12-day buffer on a SKU selling 5.2/day is 63 units. It needs no supplier data, which makes it the default first method, but it's blind to how long replenishment actually takes.

Is it acceptable to guess safety stock? Yes, with a correction loop attached. Start at two weeks of demand, review the SKU weekly, and adjust after every bare cycle or near miss. The guess converges on a defensible number within a season or two. What's not acceptable is guessing once and never revisiting — an unreviewed guess drifts as the supplier and demand change.

When can I stop using a fallback and use the real formula? After roughly six POs per supplier with recorded order and receive dates, plus enough sales history to compute daily demand variability. At that point measured lead time and measured variability beat every proxy. Until then, keep the fallback running and record the dates — the history is what makes the upgrade possible.

What if the supplier's quote turns out to be wrong? Replace it with the measured span from your first delivery — that comparison is the whole point of recording dates. Until the correction lands, widen the padding rather than trusting the original quote, because the quote's most common flaw is covering production while omitting freight, customs, and your own receiving time.

Next step

Which three SKUs are you reordering with no lead-time data at all? Pull daily demand tonight for three of them, run each through the triage table, and write the fallback buffer under its reorder line. Then start recording order and receive dates on the next PO so the upgrade clock starts — the real formula they'll eventually feed is in the full safety stock math — and set a monthly reminder to revisit the buffers until measured history retires them.