Most e-commerce brands pick their first two warehouse locations based on conventional wisdom — Los Angeles, then New Jersey. That logic is reasonable. But deciding where to go next requires something conventional wisdom can't provide: your own fulfillment data.
The pattern is almost universal among e-commerce sellers sourcing from Asia. The first overseas warehouse goes in Los Angeles — low ocean freight costs, abundant 3PL options, a well-established logistics ecosystem. As volume grows and the limitations of a single West Coast node become apparent, the second warehouse goes in New Jersey, capturing the roughly 70% of US population and purchasing power that sits east of the Rockies.
Neither of these decisions is wrong. They're sensible defaults, and for most brands at an early stage, they hold up. But the third warehouse is a different kind of decision — one where defaults are no longer good enough, and where the right answer varies significantly from brand to brand.
This post shares the analytical framework we used with a well-established overseas warehousing operator to help them plan their warehouse expansion roadmap — and the surprisingly costly inventory placement patterns we uncovered along the way.
When we were engaged to help plan this operator's warehouse expansion strategy, the temptation was to approach it as a real estate question: which markets have the best capacity, cost, and labor availability? But that frames the problem from the wrong direction.
The right starting point is the customer's core requirement: get my products to consumers as fast as possible, at the lowest possible cost. In fulfillment terms, this is an inventory placement problem — keeping stock geographically close to demand. Before you can decide where to add warehouse capacity, you need to understand where your customers' actual demand is, and whether their current inventory is positioned to serve it efficiently.
We asked the operator to share fulfillment data from their key accounts. What we found was striking.
We classified every outbound shipment across three major accounts into five placement categories, based on the relationship between the shipping origin and the delivery destination:
Goods shipped to a Zone 2–3 address within the same region. Delivery in 1–2 days. Ideal consumer experience at the lowest unit freight cost.
Goods shipped to an adjacent region. Delivery in 2–4 days. Still a strong consumer experience with manageable freight cost.
Goods that could ship from either LA or NJ, taking 5–10 days regardless. Shipping from LA carries a lower inbound freight cost to the warehouse, making it marginally preferable.
Same delivery window as Category 3 (5–10 days), but shipped from the NJ node at higher inbound cost. Avoidable with better inventory placement.
Goods shipped across the entire country. Worst-case: Category 5b — inbound stock sent from Asia to the NJ warehouse, then outbound orders fulfilled to West Coast consumers. A complete U-turn: maximum freight cost, maximum delivery time, minimum consumer satisfaction.
After analyzing shipment data across the operator's three largest accounts, the breakdown was sobering:
| Account | Total CBM | Cat 1 ✅ | Cat 2 ✅ | Cat 3 ⚠️ | Cat 4 ⚠️ | Cat 5 ❌ |
|---|---|---|---|---|---|---|
| Account A | 32,955 | 39% | 17% | 11% | 7% | 26% |
| Account B | 13,741 | 32% | 18% | 10% | 6% | 33% |
| Account C | 37,689 | 40% | 18% | 13% | 7% | 21% |
| Total / Average | 84,384 | 39% | 18% | 12% | 7% | 25% |
FEU equivalent: ~1,241 containers. Category 4+5 combined: 32% of total volume.
💡 The headline finding: Across these three accounts, 32% of total outbound volume fell into Categories 4 or 5 — inefficient placement patterns that generate excess freight cost, longer delivery times, and worse consumer experience. This isn't a marginal problem. It represents nearly one-third of all shipments operating below the standard that the current two-node network could theoretically support with better planning.
The consequences of poor placement extend beyond the brand owner. The 3PL operator also bears indirect costs: handling out-of-region freight adds operational complexity, creates capacity distortion when expanding warehouse footprint, and can mislead expansion planning if the underlying volume patterns aren't understood.
To quantify the opportunity, we modeled a single SKU across one of the accounts — tracing how inventory and freight metrics change as the network expands from 2 warehouse nodes to 5.
By redistributing stock across five strategically located nodes — with inventory sent to each location in smaller, more frequent replenishment quantities based on regional demand — two things happen simultaneously. Total replenishment volume decreases (because less buffer stock is needed to hedge against cross-country shipping delays). And average inventory holding drops from 7.5 months to 4.7 months of supply.
The combined financial impact — lower procurement volume, reduced inventory carrying costs, and meaningfully lower per-unit outbound freight — is substantial at scale.
Based on the full dataset analysis across all target accounts, our recommendation was a phased build-out of five warehouse nodes to meet current inventory placement needs within budget. The optimal sequence, ranked by impact per dollar of warehouse cost, was:
| Step | Avg Delivery Days | Warehouse Network | |||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | 4.6 days | LAX | |||||||
| 2 | 4.4 days | LAX | EWR | ||||||
| 3 | 4.1 days | LAX | EWR | SAV | |||||
| 4 | 3.9 days | LAX | EWR | SAV | DAL | ||||
| 5 | 3.2 days | LAX | EWR | SAV | DAL | CHI | |||
| 6 | 2.9 days | LAX | EWR | SAV | DAL | CHI | OAK | ||
| 7 | 2.6 days | LAX | EWR | SAV | DAL | CHI | OAK | MIA | |
| 8 | 2.0 days | LAX | EWR | SAV | DAL | CHI | OAK | MIA | SEA |
Node codes: LAX = Los Angeles, EWR = Newark/New Jersey, SAV = Savannah, DAL = Dallas, CHI = Chicago, OAK = Oakland, MIA = Miami, SEA = Seattle. Sequence derived from fulfillment data analysis; optimal nodes will vary by brand.
The warehouse expansion analysis was conducted for a 3PL operator, but the underlying logic applies directly to brand owners making their own fulfillment network decisions.
A few principles worth internalizing:
The analysis behind this work used a deliberately simple stack. Conceptual modeling and dynamic scenario planning (warehouse layout, inventory impact by node) was done in Excel. Large-scale shipment data processing was handled in SQL. Results were visualized back in Excel using US map heatmaps to make geographic patterns immediately interpretable to non-technical stakeholders.
The US heatmap Excel template used in this analysis is available to share. If you're interested in applying this framework to your own fulfillment network, reach out to us directly.