Why "pick one fulfillment method" stops working
Early on, most sellers pick a single fulfillment approach and apply it to everything — usually because there's only one or two SKUs and one or two channels to worry about. That approach breaks down as the catalog and channel mix grow: a 40-SKU catalog selling on Amazon, Walmart, and a DTC site almost never has one fulfillment method that's simultaneously cheapest, fastest, and most reliable for every single SKU on every single channel. Established multichannel sellers instead build a fulfillment mix — different SKUs, or the same SKU across different channels, deliberately routed to different fulfillment methods based on their individual economics.
The core dimensions that drive the split
- SKU velocity — fast movers often justify the fee premium of marketplace fulfillment for the badge/ranking benefit; slow movers accumulate expensive long-term storage fees under FBA/WFS and are often better served by a 3PL or self-fulfillment with a smaller on-hand quantity.
- Size and weight tier — small, light SKUs are usually cheapest and most effective under FBA/WFS; bulky or heavy SKUs frequently flip the math toward a 3PL (see FBA vs. WFS vs. 3PL).
- Channel mix per SKU — a SKU sold only on Amazon has a simpler answer than one sold identically across Amazon, Walmart, and your own site, where pooling inventory in one location (usually a 3PL, or FBA with Multi-Channel Fulfillment) can reduce total carrying cost.
- Margin sensitivity — a thin-margin SKU is more sensitive to a fulfillment fee difference than a high-margin one; run the actual contribution margin (see The Full Cost Stack of a Marketplace Sale) before deciding, not just top-line fee comparison.
- Special handling needs — hazmat, temperature-sensitive, or fragile items may simply not be viable under a given fulfillment program regardless of cost, which narrows the real options before cost even enters the decision.
A worked example: three SKUs, three answers
| SKU | Profile | Likely best fit |
|---|---|---|
| Small phone accessory | Light, fast-moving, sold on Amazon and a DTC site | FBA for Amazon orders (badge benefit); Multi-Channel Fulfillment or a 3PL for the DTC site, depending on order volume there |
| Bulky home good | Heavy, moderate velocity, sold on Amazon and Walmart | 3PL for both channels, pooling inventory rather than splitting stock across FBA and WFS separately |
| Seasonal decor item | Light but highly seasonal, sold only on Etsy and DTC | Self-fulfillment or a smaller 3PL allocation, since marketplace fulfillment's storage fees would be expensive to carry for months of low demand between seasons |
The pattern across all three: the decision is made per-SKU based on its own economics and channel footprint, not applied uniformly to the whole catalog.
Operational requirements a mixed strategy adds
Running a fulfillment mix is genuinely more complex than a single-method approach, and that complexity has to be managed deliberately:
- An order management system (OMS) or equivalent routing logic that sends each order to the correct fulfillment location based on channel and SKU — manual routing breaks down quickly past a handful of SKUs.
- Inventory visibility across every location so you're not double-committing the same unit to two channels, or missing a stockout at one location while stock sits idle at another.
- A clear allocation rule for split inventory — for example, keeping a defined percentage of a SKU's total stock in FBA and the rest at a 3PL, with a process for rebalancing as sales patterns shift.
- Reconciliation discipline — with inventory and fees split across multiple systems, monthly reconciliation (units, fees, and discrepancies) takes real, scheduled effort rather than happening automatically.
When to introduce a mix (and when not to)
A fulfillment mix is worth the added complexity once you're carrying enough SKU and channel diversity that a single method is clearly leaving money or performance on the table for a meaningful share of the catalog — commonly once a seller is selling three or more channels, or once a specific subset of SKUs (oversized items, slow movers) is visibly expensive under the current default. Below that point, the operational overhead of managing multiple fulfillment relationships usually outweighs the savings; keep it simple until the catalog and channel footprint genuinely justify the added complexity.
Common mistakes
- Splitting every SKU across every possible fulfillment method "just in case," which multiplies operational overhead without a real cost or performance justification.
- Not revisiting the split as sales velocity shifts — a SKU that was a fast mover under FBA six months ago may have slowed enough that it's now accumulating long-term storage fees.
- Underinvesting in order routing and inventory visibility tooling, leading to stockouts at one location while stock sits idle at another.
- Treating the fulfillment mix as a one-time project rather than an ongoing operational discipline that needs periodic review.