Inventory Forecasting: How Not to Run Out (or Overstock)

There are two ways inventory quietly kills a small brand, and they pull in opposite directions. Run out, and you lose sales, momentum, and ranking right when things are working. Buy too much, and you tie up cash in boxes you can't sell fast enough — paying storage fees the whole time. Most new sellers guess their way between those two cliffs. Forecasting is how you stop guessing.

We've covered how to send inventory to Amazon and the MCF setup that fulfills your store orders — but those guides assume you already know how much to send and when. This one fills that gap: sales velocity, lead time, reorder points, and safety stock, plus the storage fees that punish overstock. None of it requires software you don't already have. A spreadsheet and four numbers will do it.

The twin costs: stockouts and overstock

Every inventory decision is a trade-off between two failures. Understanding what each one actually costs is the whole game.

A stockout kills more than the missed sale

When you run out, the obvious loss is the revenue you didn't earn while the listing said "unavailable." But the real damage is downstream. On Amazon, sales velocity feeds your Best Sellers Rank and your organic position — go dark for a week and the algorithm hands your slot to competitors who stayed in stock. You also lose the Buy Box, because you can't win a box on a product you can't ship. When you finally restock, you're not picking up where you left off — you're climbing back from a lower rung, often spending on ads just to rebuild the visibility you used to get for free. A stockout doesn't cost you a few days of sales; it costs you the weeks of momentum it takes to recover.

Overstock bleeds margin you never see

Overstock feels safer — you'll never miss a sale — so sellers over-correct into it constantly. But every unit sitting in a warehouse is cash you can't reinvest, and it's costing you rent. Amazon charges monthly storage on everything you hold, and that rate climbs in Q4 and again once inventory ages past certain thresholds. Buy a year of stock to feel secure and you've frozen your working capital and signed up to pay aged-inventory surcharges on the slow-movers. The cost is invisible on your P&L until you add it up — which is exactly why it's dangerous.

Sales velocity and days of cover

Everything downstream is built on two numbers, so get these right first.

Sales velocity is your average units sold per day. Pull your last 30 to 60 days of unit sales and divide by the number of days. If you sold 360 units in the last 30 days, your velocity is 12 units/day. Use a recent, representative window — not a launch spike, not a dead patch — and recompute it regularly, because velocity is the input every other calculation depends on.

Days of cover flips that around: how many days will your current stock last at the current velocity? Divide units on hand by daily velocity. 480 units at 12/day is 40 days of cover. This is the single most useful number on your whole inventory dashboard, because it answers the only question that matters in plain language — how long until I run out? Once you know your days of cover and your lead time, you know whether it's time to reorder.

Lead time is longer than you think

Lead time is the total clock from "I place the reorder" to "those units are live and sellable." New sellers chronically underestimate it because they only count one piece. It's actually three stacked stages:

• Production — your supplier's manufacturing time once the order and deposit are in. This is the lead time you negotiated when vetting your supplier, and it can run weeks to months.
• Transit — freight from the factory to the fulfillment center. Ocean freight from overseas is slow and variable; air is faster and pricier. Customs can add unpredictable days on top.
• Receiving / check-in — the part everyone forgets. Once your shipment arrives at Amazon, it isn't sellable until it's checked in and distributed, which can take additional days or longer in peak season.

Add all three and use the realistic total, not the best case. If production is 25 days, transit is 20, and receiving is 10, your lead time is 55 days — not the "about a month" your supplier mentioned. Track what your reorders actually take and use real numbers. A lead time you wish were true is how stockouts happen.

The reorder point: your trigger line

Here's where the numbers come together. The reorder point is the stock level that should trigger your next order. Hit it, and you place the reorder; stock keeps draining during the lead time, and your new units should land just as you're about to run dry. The formula is simple:

Reorder point = (Avg daily sales × Lead time days) + Safety stock. The first part covers normal demand across the wait: at 12 units/day over a 30-day lead time, you'll sell about 360 units before the restock arrives, so you need at least that much on hand when you order. Add 80 units of safety stock for the things that go wrong, and your reorder point is 440. The instant on-hand inventory touches 440, you place the order — no debate, no "let's see how next week goes." That discipline is the entire point: the line removes the judgment call that's usually made too late.

Safety stock: buffer without bloat

If forecasts were perfect, you'd need no buffer — order at exactly velocity-times-lead-time and coast to zero as the restock lands. But forecasts aren't perfect: demand spikes, a freight delay tacks on a week, the factory slips. Safety stock is the cushion that absorbs that variability so a single bad week doesn't become a stockout.

The trick is sizing it. Too little and the buffer does nothing; too much and you're back to overstock and storage fees. Two factors should push your safety stock up: demand variability (if your daily sales swing wildly, you need more buffer than a steady seller) and supply unreliability (a supplier who's late half the time forces a bigger cushion than a metronome). A practical starting point is to cover the gap between your typical and your worst-case lead time — if reorders usually take 30 days but occasionally 45, hold roughly the extra demand those 15 days represent. Tune it from there as you gather real data. Safety stock is insurance: you want enough to cover realistic trouble, not every conceivable disaster.

Seasonality and promo spikes: forecast up, not flat

A flat forecast — "we sell 12 a day, so we'll always sell 12 a day" — is the most common forecasting mistake, and it guarantees a stockout right when demand matters most. Real demand isn't a straight line. It climbs into Q4 and gift seasons, jumps the moment you run a promotion or a coupon, and spikes if a post or an influencer placement lands. Your reorder before a known peak has to be sized for the peak, not for the quiet average that came before it.

That means looking ahead. If you know a holiday surge or a sale is coming, raise your velocity assumption for the weeks the reorder will cover, and place that order earlier than usual to clear the longer lead times that plague peak season. Our guide to launch timing and seasonality digs into the calendar; the inventory takeaway is simple — forecast up into a spike, and order ahead of it, because the worst time to be late on stock is the best time to be selling.

Amazon storage and aged-inventory fees

This is the cost that makes overstock more than a cash problem. Amazon charges monthly storage fees on every unit you keep in their warehouses, billed by the volume your inventory occupies. That rate isn't flat — it rises sharply in the Q4 peak, so the slow-movers you over-ordered cost the most to store exactly when space is tightest.

On top of that sit aged-inventory surcharges: extra fees on stock that's been sitting too long. Inventory that overstays gets hit with escalating charges designed to push you to either sell it or pull it out. Buy a year of cover to feel safe and the math can flip — the storage and aging fees on the back half of that order quietly eat the margin you thought you were protecting. This is why "just order extra to be safe" is bad advice past a point. See how Amazon seller fees work for the full breakdown, and model a reorder's landed-plus-storage cost with the FBA & MCF calculator before you commit to a big batch. The goal isn't maximum stock — it's enough cover to never run out, and not a box more.

A simple reorder worksheet

You don't need a forecasting platform. Build one row per product in a spreadsheet with these columns and update it weekly:

• Avg daily sales — units sold in your last 30–60 days ÷ days in that window.
• Lead time (days) — your real production + transit + receiving total, not the optimistic one.
• Safety stock — your buffer for demand and supply variability.
• Reorder point — (avg daily sales × lead time) + safety stock.
• Units on hand — current sellable inventory.
• Days of cover — units on hand ÷ avg daily sales.
• Reorder qty — how much to bring back (cover your next cycle plus any upcoming spike, without overshooting into storage-fee territory).

Then add one flag column: order now? — a simple check of whether units on hand has dropped to or below the reorder point. When it flips to yes, you order. That's the entire system. Recompute velocity regularly, bump your forecast ahead of known peaks, and the spreadsheet tells you when to act before you're staring at an empty shelf.

Make it a habit, not a panic

Forecasting isn't a one-time setup; it's a weekly five-minute check. Velocity drifts, lead times shift, seasons turn — the numbers that were right last month quietly go stale. The sellers who never stock out and never drown in dead inventory aren't lucky or psychic; they just look at the same four numbers every week and let the reorder point make the call. If you're still mapping out the fundamentals, start with FBA for beginners and sending your inventory in, then come back and put a reorder point on every SKU you sell — including, if you're us, both Zubiflex 10-packs.