How much stock do you actually need for a flash sale?
How much inventory for a flash sale? A sizing method that multiplies your baseline by a measured uplift, plus the sell-down plan for the week after.
InventoryX is in development and not available today. ClearStock is the live Shopify app.
Short answer: size a flash sale as baseline daily velocity × sale days × your measured uplift multiplier, plus one bad day of buffer — then check the answer against days of cover so the promo doesn't drain the shelf you need for the week after. For a phone stand selling 22/day with a measured 2.6× uplift at 20% off, a 3-day sale needs about 229 units on the shelf when it opens — and the post-sale trough, not the sale itself, is where most stores actually go dark. Here's the worksheet, and the sell-down plan that finishes the job.
The three numbers behind the answer
Every flash-sale sizing problem is three inputs, not one:
- Baseline velocity — trailing 30 days, in-stock days only. If the SKU was dark for part of the month, the raw average lies; exclude the dark days.
- The window — sale days. Fixed by your plan.
- The uplift multiplier — yours, measured from your own last promo: its daily sales ÷ the trailing baseline at the time. There is no universal multiplier; a 20% code to your email list and a 20% code to cold ads produce different animals.
If you've never run a promo, your first flash sale is a measurement, not a revenue event: run it short (48 hours), cap the units, and read the multiplier afterward. Estimating demand with no history is its own problem — the zero-history forecasting methods apply here too.
The sizing worksheet, worked
Illustrative numbers on a realistic catalog line: the aluminum phone stand, STND-ALU-BLK, $19.90 retail, $4.20 illustrative unit cost, 20% off planned.
| Line | Value | Where it comes from |
|---|---|---|
| Baseline velocity | 22/day | Trailing 30 days, no dark days |
| Window | 3 days | Your promo plan |
| Uplift multiplier | 2.6× | Last 20%-off event did 57/day: 57 ÷ 22 ≈ 2.6 |
| Window demand | 172 units | 22 × 3 × 2.6 |
| Window buffer | +57 units | One bad day at promo rate |
| Stock needed on sale morning | 229 units | 172 + 57 |
| Pre-sale drain | 220 units | 22 × 10 days until the sale opens |
| On hand today | 480 units | Your position |
| Sale-morning position | 260 units | 480 − 220 ✓ covers 229 |
| After the sale | 31 units | 260 − 229 |
| Post-sale trough demand | ~13/day | Buyers pulled purchases forward; assume ~60% of baseline for the first week |
| PO placed today (19-day lead) lands day 19 | 78 units needed | Days 14–19 at 13/day, against 31 on the shelf |
That last row is the whole point of the worksheet. The sale itself is covered. But right after your biggest marketing moment — days 14 through 19, before the replenishment PO lands — you're about 2 dark days, roughly 25 units short. The fix costs almost nothing: add 60 units to the PO you place today (60 × $4.20 = $252 estimated) and the trough goes from dark to comfortable. This is also why the flash-sale PO is placed weeks before the sale, not while the sale is running — a PO placed on sale day lands after the trough it was meant to fix.
If the math says you can't cover it
Three dials, cheapest first:
- Shorten the window. Two days instead of three: 22 × 2 × 2.6 = 115 + 57 buffer = 172 units needed on sale morning — check that against your projected position.
- Shallow the discount. If you've measured a 2.2× at 10% off and a 2.6× at 20% off, the shallower tier needs 145 + 45 = 190. Only trust tiers you've actually measured.
- Cap the units. Ring-fence a sale batch and let it sell out honestly. The mechanics of setting stock aside — unavailable quantities, timing, the manual discipline involved — live in the reservation guide.
One caveat for sales longer than three days: measured uplifts decay. Day one captures the pent-up demand the announcement created; by day three you're selling to whoever's left, and the daily rate typically trails the day-one peak by a wide margin. So size a five-day sale on the multiplier but hold back the last tranche — the buffer you didn't commit on day one becomes your day-four restock decision, made with two days of real sales data instead of one guess stretched across the week.
What not to do: flip "continue selling when out of stock" for the promo and sort out the oversell later. Promo oversells arrive as apology emails and refunds, and the units you oversold were needed for the trough; the overselling mechanics — and when a store might allow it on purpose — are covered properly there.
The sell-down plan: the week after is part of the size
- Plan for the trough. Purchases pulled forward mean a quiet 7–10 days. Don't schedule another promo into it and don't panic-reorder at trough velocity — the baseline recovers.
- During the sale: set per-customer caps on the hero SKU so one buyer doesn't take the buffer with them, and check remaining cover against remaining window once a day.
- After: return to full price promptly. The demand you deferred in the trough is your next two weeks of baseline — it's already paid for the moment it comes back.
- If your "flash sale" is BFCM: the math is identical, but the dates, order-by deadlines, and the countdown discipline are their own exercise — the BFCM planning guide owns that calendar.
One note from the tooling side: the worksheet needs three inputs — in-stock-only velocity, days of cover, and an honest lead time. That's the same set our Monday Buying List computes in InventoryX; it won't predict your uplift multiplier (nothing reliably will), but it keeps the baseline honest so the multiplier has something true to multiply.
Common questions
How much extra inventory do I need for a flash sale? Baseline daily velocity × sale days × your measured uplift multiplier, plus one bad day at promo rate. In the worked example: 22 × 3 × 2.6 + 57 = 229 units on sale morning — versus 66 for the same three days at baseline, which is why "add a bit extra" undershoots so reliably.
What uplift multiplier should I plan for? The one your store has measured: your last promo's daily sales divided by the trailing baseline at the time. If you've never run one, make the first sale short and capped so the multiplier it produces is safe to act on next time.
Should I restock before or during a flash sale? Before — with the PO sized to cover the sale window and the post-sale trough. A replenishment placed mid-sale arrives after the trough it was meant to prevent; in the worked example, the difference was 60 units added to a PO that had to be placed three weeks before the sale opened.
What happens if the flash sale sells more than planned? With caps and a ring-fenced batch, it sells out and you capture waitlist demand with honest dates. Without them, you either oversell or go dark mid-promo. Resist the continue-selling toggle as a promo strategy — deferred angry emails are not a sell-down plan.
How far in advance should flash-sale stock be ordered? Lead time plus about a week of margin. On a 19-day supplier lead, the promo PO goes out roughly five weeks before the sale — which means the sizing decision happens a month before the discount code ever gets written.
Next step
Before your next promo, compute your own multiplier: last event's daily sales ÷ the 30-day baseline that preceded it. Write it next to each discount depth you've ever run. That one number turns every future flash-sale sizing worksheet into a five-minute job — and it's the only uplift figure in this article you should trust for your store.