Reorder point and days of cover for Amazon FBA

The four numbers

Every replenishment decision for an FBA product comes down to four numbers, for one SKU at one fulfilment centre:

Scroll the table sideways for the remaining columns.

Everything else is arithmetic on those four. Note that days of cover uses only what is sellable now, while the reorder point and the send quantity use the stock position: sellable units plus units on the way.

daily velocity   = units sold in window ÷ days the SKU was in stock in window
days of cover    = sellable units on hand ÷ daily velocity
stock position   = sellable units on hand + units already on the way
reorder point    = daily velocity × (lead time + safety days)
create a shipment when stock position ≤ reorder point
units to send    = daily velocity × (lead time + target cover days)
                   − stock position      (send nothing if this is zero or less)

Every number in the worked examples below is illustrative, chosen so the sums are easy to follow.

Days of cover

Days of cover answers "how long until this runs out?" A SKU with 60 sellable units that sells 4 a day has 15 days of cover. It is the single most useful number on a stock report, because it puts a fast seller with 200 units and a slow one with 20 on the same scale.

One edge case matters. A SKU that sold nothing in the window has no days of cover - not infinite cover. Dividing by zero sales and reporting "999 days" makes a dead SKU look like your healthiest one. Treat zero-sales SKUs as their own list: they are candidates for removal, not for replenishment.

The reorder point

The reorder point is the stock level at which you create the next shipment. It has to cover what you will sell while that shipment is on its way, plus a buffer for the days when sales run hot or the shipment runs late:

reorder point = velocity × (lead time + safety days)
              = 4 × (14 + 5)
              = 76 units

Compare stock plus units on the way (the stock position) with the reorder point, not stock alone. Otherwise you will create a second shipment while the first is still in transit.

When inbound stock counts, and when it does not

• For days of cover, never. Units that have not been received and made sellable cannot be sold, so they do not extend how long today's sellable stock lasts.
• For the reorder decision, yes, if it is really coming. Count a shipment that has physically left you and is addressed to this fulfilment centre. Do not count one that is only a plan, one you have cancelled, or one that has sat unreceived far longer than your normal lead time.
• Per centre. Count a shipment against the centre it is addressed to, and check where it is actually received. If a shipment ends up at a different centre than you planned, move its units to that centre in your sums.
• Expect some shortfall. Units can be received short (see the inbound discrepancies guide). If a SKU often receives short, count a little under what you shipped.

Safety days, and variable demand

Safety days are a flat buffer, which is right when sales are steady. When daily sales swing a lot, a common planning method is to size the buffer from the swing instead:

safety units = z × σ × √(lead time in days)
             = 1.65 × 2 × √14
             ≈ 12 units   (about 3 days at 4 a day)

σ is the standard deviation of units sold per day over your window (2 in this illustration) and z is a service-level factor: 1.65 is roughly a 95% chance of not running out during the lead time. This is a standard inventory-planning formula, not an Amazon rule. It assumes daily sales are roughly independent and the lead time is steady; if your lead time also varies, use a larger buffer.

Use that figure in place of safety days × velocity. The reorder point then becomes velocity × lead time + safety units.

How much to send

The reorder point tells you when. The quantity comes from how many days of stock you want to have when the shipment lands - the target cover. You will keep selling while it travels, so the lead time goes in too:

units to send = velocity × (lead time + target cover)
                − (on hand + on the way)
              = 4 × (14 + 30) − (76 + 0)
              = 100 units

Sent at the reorder point in the chart above, those 100 units land when 20 are left, which takes stock to 120 - exactly 30 days of cover at 4 a day.

The target is a trade-off. Longer cover means fewer, larger shipments and less risk of running out; it also means more stock sitting in the fulfilment centre collecting storage fees (Amazon adds a long-term fee after 150 days), and more exposed if demand drops. Unilytic's send list defaults to 30 days of cover and flags stock as excess beyond 45 - starting points to tune, not rules. One difference to know: it counts that target from today rather than from arrival, and recalculates every time your ledger syncs, so it suits topping up little and often. If you ship less often, set a longer target.

Running out costs you sales you never get back. Overstocking costs you rent.

Getting velocity right

Velocity is where most replenishment maths quietly goes wrong.

Scroll the table sideways for the remaining columns.

The formulas also assume velocity stays the same through the lead time and the target cover. If demand is seasonal or a sale is coming, use the velocity you expect over that period instead of last month's: the same weeks last year, adjusted for growth, is a common starting point.

Per fulfilment centre, not per country

FBA stock in India does not sit in one place. In one seller's ledger, stock and sales touched 31 different fulfilment-centre codes in three months. A national total can show a comfortable 40 days of cover while the centre serving one region is empty.

So run every calculation above per centre: velocity from that centre's shipments, stock from that centre's sellable balance, and a send quantity for that centre. The Inventory Ledger guide explains how to pull stock and sales per centre from the Summary view.

Measuring your lead time

Lead time is the number most sellers guess, and the one that moves the reorder point most. Measure it from your own last several shipments: the day you created the shipment to the day its units became sellable. It is made of:

1. packing and labelling at your end,
2. transit to the fulfilment centre,
3. waiting for the centre to receive the shipment, and
4. receiving and checking the units in until they show as sellable.

Use a typical recent figure, not your best one. The last two steps are outside your control and vary by centre and by season, so a lead time that was right in a quiet month can be a week short in a busy one. Treat the figure as a planning estimate from your own history: this guide does not rely on any receiving time from Amazon, and none should be assumed to be guaranteed. Unilytic's Ship Now screen assumes 14 days until you change it.

The send list, worked out for you

Unilytic runs this arithmetic on your Inventory Ledger each time it syncs, per fulfilment centre, and turns it into a ranked send list: which centres to ship to first, which SKUs, and how many units - counting stock already on the way.

app.unilytic.in/ship-nowFull size

Questions sellers ask