Real lead time: why your restock always arrives late

It happened again. You placed the restock order “with plenty of time,” you reminded yourself that the supplier takes about three weeks, and the product still landed six days after you’d already run out. You lost sales, you lost ranking on the listing, and worst of all: you’re fairly sure you ordered on time. The problem wasn’t your memory. The problem is that the number you planned with —that rounded “three weeks”— is not your real lead time.

Real lead time is not what your supplier promises. It’s the entire span from the moment you decide to restock until a unit is sellable to your customer. And between those two points there’s a chain of steps almost nobody measures end to end: production, transit, customs, receiving at the 3PL warehouse or Amazon inbound, the time MercadoLibre takes to make the stock live, and the dead days you add yourself without noticing (the weekend nobody processed the order, the holiday, Monday’s inventory cutoff).

When you run a single SKU in a single marketplace, you can carry all of that in your head. When you run dozens of SKUs across Amazon Seller Central, MercadoLibre, and a 3PL, each on its own clock, your head can’t keep up. You end up estimating with one generic number for everything, and that generic number is exactly why your restock always arrives late.

what counts as lead time and what you forget to measure

Ask ten sellers what their lead time is and nine will tell you how long the supplier takes to ship. That’s the first mistake. The lead time that matters for not stocking out is end to end, and it’s made of segments that happen in different places under different owners.

There’s the order processing time (from when you send it to when the supplier confirms and starts). There’s production or picking at origin. There’s transit, which on imports from Asia can swing by weeks depending on the carrier and the season. There’s customs, which in Mexico is not a fixed number: a pedimento that gets stuck adds unpredictable days. There’s receiving at your 3PL or inbound at an Amazon fulfillment center, where the product has physically arrived but isn’t sellable yet. And there’s the silent last segment: the time each channel takes to reflect that stock as available.

Each of those segments has its own variability. Collapsing them into a single average number hides exactly what you need to see: where the time goes and how reliable each stage is.

Dictionary: a forecast projects your future demand; without a correct real lead time, even the best forecast orders late.

inbound and receiving: the invisible days

This is the segment most sellers ignore and the one that charges the most. Your order “arrived” —the truck unloaded at the warehouse— but the unit still can’t be sold. In Amazon FBA, inbound can take several days between the center receiving the box and the unit showing as available, especially in peak season. At a 3PL, it depends on that warehouse’s receiving queue: if you landed on a Friday, they might process you Monday or Tuesday.

If your planning says “it arrives on day 21 and I can sell that day,” you’re counting wrong. The physical arrival date and the real availability date can be three, four, or five days apart. Multiply those days by your daily sell-through and there you have, precisely, the units you’ll lose to a segment that wasn’t even on your radar.

That’s why real lead time has to end at “sellable,” not at “received.” It’s the only definition that protects you from a stockout.

the average trap and variability

Say your last five restocks took 24, 26, 22, 38, and 25 days. The average is 27. If you plan with 27, you’ll stock out every time the 38-day episode repeats, and that episode isn’t rare: it’s your variability tail, and that’s where stockouts live.

Planning with the average is planning to be right half the time. What you actually need is to plan with a high percentile of your lead time —day 38, not day 27— or, better, to translate that variability into safety stock. The more unstable a segment is (typically customs and transit), the more buffer you need, not because you’re a pessimist, but because the cost of stocking out is almost always higher than the cost of carrying a few extra units.

The catch is that to do this you need the real history of each restock, by SKU and by channel. And that history doesn’t live anywhere clean: it lives split across emails with your supplier, your 3PL portal, and the inbound dates in Seller Central.

why the problem gets worse when you’re multichannel

The same product can have three different real lead times at once. The unit you send to FBA carries Amazon’s inbound. The one you leave at your 3PL to fulfill MercadoLibre Full carries the 3PL receiving plus the shipment to a MELI center. The one you sell from your own warehouse has yet another clock. Same SKU, three clocks.

When each channel lives in its own dashboard, pulling this together by hand is a half-day job that always ends in a spreadsheet with yesterday’s data. And deciding a restock with yesterday’s data, on an average that hides the variability, across three channels at once, is exactly the recipe for “I ordered on time and it still arrived late.” The problem isn’t that you’re disorganized; it’s that you’re rebuilding by hand a truth that changes every hour.

This is where a single real-time source of truth changes the conversation. If the system knows your sell-through per channel and the real lead time measured by segment, the reorder point stops being a hunch. This connects directly with why spreadsheet forecasting breaks at 200 SKUs: it’s not that the spreadsheet computes wrong, it’s that it can’t stay alive with that many clocks at once.

Dictionary: the reorder point is the stock level that triggers the order; build it on real lead time, not the promised one.

how to measure your real lead time, step by step

You don’t need a system to start measuring; you need to record four dates for every restock: when you issued the order, when it physically arrived, when it became sellable, and how many units you ordered. With that you can already compute end-to-end real lead time and compare it against what you assumed.

Do this by SKU and by channel for three or four cycles. You’ll discover two things. First: your real lead time is almost always longer than your estimate, usually because of the inbound or receiving you weren’t counting. Second: the variability concentrates in one or two segments, not all of them. Attacking those segments —switching carriers, sending to inbound earlier, negotiating receiving priority with your 3PL— pays off more than any formula tweak.

The reorder point then becomes very concrete: daily sell-through per channel, multiplied by your real lead time at a high percentile, plus the safety stock that justifies that SKU’s variability. When sellable availability touches that number, you order. Not earlier, freezing capital, nor later, giving away sales.

the signal that really matters: days of inventory

After measuring lead time, the metric you have to watch every day isn’t “how many units are left,” but “how many days of sales are left at the current pace.” That’s the only figure you can honestly compare against your real lead time.

The rule is simple: if a SKU’s days of inventory fall below its real lead time, you’re already late, no matter how many units are in the box. A product with 400 units and a 45-day lead time selling 12 a day has barely 33 days of coverage: you should have ordered already. That comparison, run automatically and in real time across all your SKUs and channels, is the difference between planning and firefighting.

Dictionary: days of inventory measure how many days of sales you have left; always compare them against your real lead time, not a fixed number.

in summary

Your restock isn’t late because of bad luck. It’s late because you plan with a lead time that’s shorter than the real one: you skip inbound and receiving, you use an average that hides the variability, and you do it across three channels that hold different clocks for the same product. Measure end to end up to “sellable,” plan with the high percentile, compare your days of inventory against that real lead time, and make that comparison live in one place that’s actually current —not in yesterday’s spreadsheet.