3PL vs in-house: rent the warehouse, own the brain.
“Should we run our own warehouse?” is the most over-romanticized question in DTC operations. Founders imagine control, speed, and savings. The numbers usually say: rent it until you can’t. Below ~500 orders a month, doing it yourself is cheaper. In the 1,000–10,000 range, a 3PL almost always wins on cost and flexibility. In-house only pulls ahead again at serious scale or genuine product weirdness. This is the real decision framework — with 2026 cost benchmarks, the major players’ current state, and the brands that brought it in-house and regretted it.
- Rough break-even is 500–1,000 orders/month: below that, in-house is cheaper; above it, 3PL economies of scale win. In-house gets painful past ~20–30 orders/day.
- 2026 3PL benchmarks: $2.50–$5.00 pick & pack, ~$20/pallet/mo storage, $25–$50/pallet receiving; all-in $5–$10 per light parcel. Logistics runs 10–15% of gross sales for most brands.
- The core trade is fixed vs variable cost: in-house fixed costs idle in slow months; 3PLs convert them to variable that scales with volume.
- Even celebrated “lean” operators mostly rent: Ridge uses ShipMonk; Gymshark uses third parties. They brought order management in-house, not the warehouse.
- Genuine in-house only makes sense at Chewy/Wayfair/Thrive scale, or for big/heavy/regulated products a 3PL handles badly.
The decision framework
Strip away the romance and it’s an arithmetic problem about fixed vs. variable cost and where your volume sits. A warehouse is a fixed-cost machine: lease, labor, WMS, racking, equipment — you pay for it whether 200 or 20,000 orders ship. A 3PL converts all of that into a per-order variable cost. At low volume, fixed costs spread over few orders make in-house brutal per unit; at very high and stable volume, the fixed-cost machine finally beats the 3PL’s margin stacked on every order. The danger zone is the middle, where founders build a warehouse for the volume they hope to have and then carry idle fixed cost through every slow month.
The 2026 cost benchmarks
| Line item | 2026 benchmark |
|---|---|
| Pick & pack | $2.50–$5.00 / order (SMB); $1.50–$3.50 standard |
| Storage | ~$20/pallet/mo; ~$0.46/cu ft/mo |
| Receiving | $25–$50/pallet or $0.25–$0.75/unit |
| Returns processing | $3–$10/return (apparel return rates 30–40%) |
| Monthly minimum | ~$517 avg (up from ~$438 in 2024); setup $500–$2,000 |
| All-in per light parcel | $5–$10 |
| In-house warehouse wage | ~$18/hr median; turnover 40–49% |
| Industrial lease | ~$10/sq ft/yr US avg ($16–$18 in LA/OC) |
Watch the hidden fees — kitting, special projects, and especially long-term storage penalties (1.5–3x standard after 30/60/90 days) and peak-season surcharges (carriers added 6–9% over 2024 in the 2025–26 peak). The quoted per-order rate is rarely the all-in rate. And note: most “average 3PL cost” figures trace back to a single 2025 survey of 600+ warehouses, so treat them as one data point, not gospel.
The major players, mid-2026
ShipBob (40+ fulfillment centers, ~100M orders/yr, last valued ~$1.1B) and ShipMonk (3M+ sq ft, hybrid pricing, pushing bonded fulfillment to cut tariffs) anchor the SMB-to-mid market. Red Stag owns the big/heavy/bulky niche with financially-backed accuracy guarantees. The cautionary tale is Shopify’s logistics retreat: Shopify bought Deliverr for ~$2.1B in 2022, then sold it plus the Shopify Fulfillment Network to Flexport in 2023 for equity (~13% of Flexport) — a multibillion-dollar admission that even Shopify couldn’t make first-party fulfillment pay. And Amazon MCF/FBA remains the gravity well, with 2026 fee increases (~3–5% plus a new ~3.5% fuel surcharge) that quietly raise everyone’s floor.
The brands famous for being lean didn’t build warehouses. They rented the warehouse and built the software — the order management, the routing, the data — that makes the rented warehouse perform.
When in-house actually wins
Three cases justify owning fulfillment. Scale: Chewy, Wayfair (15 warehouses, 95% of the US in two days), and Thrive Market run their own because the fixed-cost machine finally beats per-order economics. Product weirdness: oversized, hazmat, cold-chain, or heavily kitted goods that 3PLs price punitively. Brand-critical experience: custom packaging or assembly that is the product. Outside those, the celebrated lean operators — Ridge among them — rent the warehouse and pour their energy into the layer that compounds: order management, distributed-inventory strategy, and the automation that cut Ridge’s CX team from 10 to 4. Warehouse robotics is real (AMR adoption up ~45% YoY, sub-24-month paybacks) but it’s mostly being deployed by the 3PLs, which is one more reason to let them carry the capex.
- Default posture across the portfolio: rent fulfillment, own the brain — order management, inventory placement, and data stay in-house; pick/pack/ship is a 3PL until scale says otherwise.
- We run distributed inventory (2–4 nodes) to buy 2-day delivery without owning a single brick — ~25–30% parcel savings at 3 nodes.
- “Bring it in-house” is a capital-allocation decision underwritten on stable volume and payback, not a founder’s pride. Most of the time, the math says rent.
Sources: Extensiv; The Fulfillment Advisor (2025 600+ warehouse survey); CNBC (Shopify/Flexport); Shopify (Ridge fulfillment + AI); Sacra (ShipBob); BLS (warehouse wages); Amazon Seller Central (2026 FBA/MCF fees, via third-party trackers).
Own a brand this applies to?
We talk to founders at every stage — long before they're ready to sell.
Start a Conversation →