Order Routing

One-Line Definition

Order Routing is the automated decision process that assigns each customer order to the single best fulfillment node — warehouse, 3PL, dark store, dropship vendor, or retail store — based on real-time inventory, destination distance, shipping cost, capacity, and service-level rules.

Real-Life Analogy

Think of order routing as air traffic control for parcels.

When you book a flight, you don't choose the runway or the gate — a control tower decides which airport, which terminal, and which slot gets your plane based on weather, congestion, and fuel cost. You just want to land on time.

Order routing does the same for e-commerce. A customer clicks "Buy." Behind that click, a routing engine evaluates dozens of possible fulfillment nodes in milliseconds and picks the one that gets the order out fastest, cheapest, and most reliably. The customer sees "Order confirmed." The system has already made 40+ micro-decisions before the confirmation page loads.

Core Formula

There is no single industry-standard equation, but most production routing engines optimize a weighted cost function like this:

Route Score = (w₁ × Shipping Cost)
            + (w₂ × Distance / Zone Factor)
            + (w₃ × Fulfillment Cost)
            + (w₄ × Split-Shipment Penalty)
            + (w₅ × SLA Breach Risk)
            − (w₆ × Inventory Availability Score)
            − (w₇ × Node Capacity Headroom)

Where weights (w₁…w₇) are tuned per merchant, per channel, and often per SKU class.

Worked example. A US merchant has three nodes that can fulfill a 2-item order to a buyer in Denver:

Node B wins — not the cheapest shipping, not the cheapest fulfillment, but the best *blended* outcome. That blend is the whole point.

Comparison with Related Terms

Key distinction: Allocation decides *what* stock is committed. Routing decides *where* the order is fulfilled. Rate shopping decides *who* carries it. They are sequential, not interchangeable.

Use Cases

1. Cross-border DTC with regional 3PLs.

A Shopify Plus brand selling into the EU holds stock in Rotterdam, Warsaw, and a UK bonded warehouse. Routing sends German orders to Rotterdam, Polish orders to Warsaw, and UK orders to the local node — cutting average transit from 6.2 days to 2.4 days and reducing duty exposure.

2. Marketplace with split-shipment avoidance.

A marketplace seller has SKU-A in California and SKU-B in New Jersey. Naive routing ships two parcels at $9.80 combined. Smart routing finds a third node holding both and ships one parcel at $7.10 — a 27% logistics saving and a better unboxing experience.

3. Peak-season capacity spillover.

During Black Friday, a primary DC hits 92% capacity. Routing automatically diverts overflow to a secondary node 300 miles away, accepting a $1.40 higher ship cost to protect a 2-day delivery promise on 18,000 orders.

4. Omnichannel ship-from-store.

A retailer with 240 stores routes online orders to the nearest store with stock when the DC is out. This converts ~11% of "out of stock" events into fulfilled orders and reduces last-mile distance by an average of 340 miles.

5. High-value / regulated goods.

Electronics or lithium-battery SKUs can only ship from hazmat-certified nodes. Routing applies hard constraints before cost optimization, ensuring compliance is never traded for savings.

Misconceptions

"Routing just picks the closest warehouse."

Distance is one input among many. A closer node with higher labor cost, lower capacity, or no hazmat certification often loses to a farther, cheaper, compliant node.

"It's the same as inventory allocation."

Allocation reserves stock. Routing selects the fulfillment node. A system can allocate perfectly and still route badly — or vice versa.

"Cheapest shipping always wins."

Cheapest shipping frequently means split shipments, slower SLA, or higher fulfillment labor. Routing optimizes *total landed cost per order*, not the carrier invoice line.

"You set it once and forget it."

Weights must shift by season, channel, SKU margin, and carrier performance. A Q4 routing profile that ignores capacity will break by December 5th.

"It only matters at scale."

Even a 3-node operation routing 200 orders/day can cut 15–20% of shipping spend with basic rules. Scale amplifies the benefit; it doesn't create it.

"AI/ML is required."

Rules-based routing handles 80% of value. ML improves edge cases — carrier reliability prediction, demand spikes, address-quality scoring — but is not a prerequisite.

Related Terms

- DOM (Distributed Order Management) — the platform category that houses routing logic

- OMS (Order Management System) — system of record for orders; often contains routing

- Inventory Allocation — upstream reservation of stock

- Ship-from-Store — a routing mode using retail locations as nodes

- Split Shipment — multi-node fulfillment of one order; a routing outcome to minimize

- SLA (Service Level Agreement) — delivery promise routing must protect

- Zone Skipping — injecting parcels into a carrier network closer to destination

- Landed Cost — product + shipping + duty + fees; the true target of routing optimization

- Node — any fulfillable location: DC, 3PL, dark store, store, dropship vendor

- Routing Rules Engine — the configurable logic layer that scores and selects nodes

Bottom line: Order routing is the invisible decision layer that turns "we have inventory somewhere" into "the right order leaves the right node at the right cost." Get it right and you cut shipping spend 10–30%, shorten delivery times, and protect margin — without changing a single product or price.