When to reserve stock per channel, when to pool, and how to measure.
A framework for Amazon FBA + DTC operators: decide which SKUs get channel-reserved inventory, which get pooled, and what KPIs prove the policy is working before it bites you on Q4.
The short answer
Reserve inventory per channel when a SKU's demand is concentrated in one channel and a stockout there carries a penalty beyond the lost sale — Amazon suppression or lost Buy Box, for instance. Pool inventory when demand is spread across channels and you can move stock between them faster than demand shifts.
Worked example
Take a SKU selling 70% through Amazon and 30% DTC, where FBA replenishment takes about three weeks. Because the channel is dominant and the penalty for going dark is a ranking loss that outlasts the stockout, that SKU is a reserve candidate. A SKU splitting 45/55 with a one-week transfer time should be pooled: reserving it would strand stock on the wrong side of a split that moves week to week.
Reserving stock has a cost people forget to count
Reserved inventory is unavailable to the channel that turns out to need it. Every unit you ring-fence for Amazon is a unit your DTC store cannot sell, so a reserve policy trades flexibility for protection and you should only pay that price where the protection is genuinely worth more.
That is why concentration matters more than volume. A SKU with demand split evenly gets little protection from reserving and loses a lot of flexibility, while a SKU that is 80% one channel gets most of the protection at almost no cost.
Measure the policy, or it silently stops being right
Three measures tell you whether allocation is working: per-channel fill rate, the number of stranded units sitting in a channel that is not selling them, and how often you had to make an emergency transfer between channels. The last is the clearest signal that the split no longer matches demand.
Review before every seasonal peak, not after it. Channel mix shifts fastest exactly when the cost of getting it wrong is highest, and a split set in June will not survive Q4 unexamined.
Frequently asked questions
Should I reserve inventory per sales channel or pool it?
Reserve when demand for a SKU is concentrated in one channel and a stockout there carries a penalty beyond the lost sale, such as an Amazon ranking or Buy Box loss. Pool when demand is spread across channels and you can transfer stock faster than the mix changes.
What is the cost of reserving stock for a channel?
Reserved units cannot be sold by any other channel. You are trading flexibility for protection, so it pays off on SKUs with concentrated demand and costs you on SKUs whose channel split moves week to week.
Which KPIs show that a channel allocation policy is working?
Per-channel fill rate, stranded units held in a channel that is not selling them, and the frequency of emergency inter-channel transfers. Rising emergency transfers mean the allocation no longer matches real demand.