How to plan a dropship channel
Planning methods divide demand into two kinds: booked commitments you have been ordered for, and forecast you generate yourself. Dropship is neither. The partner orders nothing, yet the demand comes from their site rather than yours.
So what you commit is not a quantity. It is an availability policy — and the failure metric is cancellation, not shortfall.
- Definition — Dropship
- Dropship is an arrangement where a retail partner sells your product on their storefront and you ship it directly to the customer. The partner holds no inventory and places no purchase order, so the commitment being made is not a quantity ordered but a quantity made available — and the risk sits with whoever exposed the stock.
- exposed units = f(live pool, channel share, feed latency, ship-window SLA)
- Used by: Planners and account managers running partner-fulfilled demand from their own inventory
- Related: Availability policy, cancel rate, virtual inventory, open-to-buy
Why existing methods do not reach it
The method for planning wholesale and DTC together rests on keeping booked commitment and forecast apart, because adding them double-counts. Dropship falls outside both categories:
- Not booked. No order exists. The partner has committed to nothing and can stop selling your item at any time.
- Not your forecast. Demand is driven by the partner’s on-site merchandising, search placement and promotional calendar. None of it appears in your history, and none of it is visible to you in advance.
In the first year especially, dropship demand is not forecastable from your own data. The honest planning object is a risk budget: how much inventory you are willing to expose to a channel whose volume you cannot predict, revised as real demand accumulates.
Feed latency and the ship window
Feed latency is how stale your availability file is by the time a customer reaches the partner’s checkout. Every minute of latency is a window in which stock can be sold elsewhere while the partner still shows it as available. A daily feed on a fast-moving item is an oversell generator.
The ship-window commitment is how quickly an accepted order must leave your warehouse. It sets how much operational buffer you need and constrains which items can be offered at all — an item with a long pick or a supplier-direct route may not be dropship-eligible regardless of how well it sells.
Cancel rate, not fill rate
In most channels, exposing too little inventory costs a sale and exposing too much costs a markdown. Dropship breaks that symmetry.
An oversold unit does not become a backorder. It becomes a cancellation — the partner’s customer is told the item is unavailable after buying it. Partners publish a cancel-rate threshold and charge back against breaches, and repeated breaches count toward suspension of the account.
So over-exposure converts into penalty and account risk, not into a delayed sale. That is why the buffer sits on the availability policy rather than on the inventory: you expose less than you hold, deliberately, and the gap is the service-level buffer.
Dynamic exposure beats a carve-out
Most brands run dropship from the same physical stock as everything else, and the instinct is to carve out a fixed number per item per channel. That number is stale the moment velocity changes — it strands units in a slow channel while a fast one sells out. The mechanism that works is dynamic exposure, sometimes called virtual reservation: expose a calculated share of live availability, recalculated as stock moves, so every channel works from one truth. Its cost is a feed frequency high enough that the calculation is still true at checkout, which loops back to feed latency as the binding constraint.
Where the units come back to
Returns on dropship orders commonly route back to you rather than into the partner’s network, though arrangements vary — some partners accept dropship returns in store and process or dispose of them themselves. Whichever applies, confirm it before planning, because it changes both the return lag and whether returned units re-enter your sellable pool at all. The returns planning method applies here with one adjustment: the shipped-to-net gap on a partner storefront is not necessarily the same as on your own, because the customer who bought there had different information.
Frequently asked questions
- How is dropship different from wholesale or your own e-commerce?
- Wholesale gives you a booked commitment: the account orders, and that quantity is yours to fulfil. Your own channel gives you a forecast you control end to end. Dropship is neither — the partner commits to nothing and places no order, but the demand is generated by their site rather than yours, so you cannot forecast it from your own history either. What you commit is not a quantity, it is an availability policy.
- What is the planning object in a dropship channel?
- An availability policy — how many units of each item you expose to the partner, and how frequently that number is refreshed. There is no receipt to plan against and no purchase order to reconcile. Two numbers govern whether the policy works: feed latency, meaning how stale the availability file is by the time a customer reaches the partner’s checkout, and your shipping-window commitment, which determines how quickly an accepted order has to leave.
- Why is cancel rate the metric that matters, not fill rate?
- Because an oversell does not become a backorder, it becomes a chargeback. If you expose stock you no longer have and the partner sells it, the resulting cancellation is counted against you and usually carries a financial penalty, with repeated breaches counting toward a suspension threshold. Over-exposure therefore converts directly into cost and account risk rather than into a delayed sale, which inverts the usual trade-off between service level and availability.
- How do you plan dropship demand in the first year?
- You do not forecast it — you set a risk budget. First-year dropship demand is driven by the partner’s on-site merchandising, search placement and promotional calendar, none of which you see in advance and none of which appears in your own sales history. So the initial exposure is a decision about how much inventory you are willing to put at risk of being sold through a channel whose volume you cannot predict, revised as actual demand data accumulates.
- How do you expose one inventory pool to several channels safely?
- With dynamic exposure against the single physical pool rather than a static per-item allocation. A fixed number carved out for each channel is stale the moment velocity changes: it strands stock in one channel while another sells out. Exposing a calculated share of live availability, recalculated as stock moves, keeps every channel working from the same truth. The trade-off is that it requires a feed frequency high enough for the calculation to still be true when the customer checks out.
- Dropship is neither a booked commitment nor your own forecast, so neither existing method reaches it.
- The planning object is an availability policy — how much you expose and how often it refreshes — not a receipt quantity.
- Feed latency and the ship-window commitment are the two numbers that decide whether the policy works.
- Cancel rate is the failure metric: an oversell is a chargeback against a suspension threshold, not a backorder.
- First-year demand is a risk budget, not a forecast — the partner’s merchandising drives it and you cannot see it.
- Expose dynamically against one live pool; a static per-channel carve-out is stale as soon as velocity changes.
Your calculator result is one number. RetailNorthstar keeps the whole plan connected — line plan, OTB, assortment, buy, POs, and production.