Same warehouse. Per-order pricing: $180,000 a year. Per-user pricing: $120,000 a year. Where does the difference come from?
That's a real selection I saw in Dallas in 2024: a 120,000-square-foot 3PL facility, 6,000 orders a day, 25 operations users, two WMS vendors quoting head to head. Vendor A charged per order, $0.12 a pop. Vendor B charged per user, $400 per user per month. Over 260 working days a year:
Vendor A: 6,000 orders x 260 days x $0.12 = $187,000 Vendor B: 25 users x $400 x 12 = $120,000
A $67,000 gap. The owner asked me which to pick. I told him not to rush toward the cheaper one — these two quotes aren't even calculated in the same dimension.
That math is wrong: one follows the weather, the other follows headcount
Per-order pricing says: your order volume is your cost. Peak season at 12,000 orders a day, slow season at 3,000, and the bill moves with volume. Fans of it call it "pay more when you earn more." Sounds fair.
Per-user pricing says: your headcount is your cost. 25 users, rock solid, $10,000 a month, $120,000 a year, and doubling order volume adds nothing.
The catch: a warehouse's order volume and headcount never move in sync. That Dallas site hired 15 temps for peak — headcount went from 25 to 40, but order volume multiplied 2.5x. Under per-user pricing, peak cost them an extra $72,000 in user fees for the year; under per-order, the peak bill more than doubled. Both models get expensive in peak season. They just get expensive differently.
The only number I look at is the order volatility ratio: your highest-volume month divided by your lowest, over the past 12 months. Under 2, the two models land close. Over 3, picking the wrong one routinely costs 30%+ extra per year. That Dallas warehouse sat at 4.2 — exactly the kind of profile where a wrong pick hurts.
The four line items hiding in the contract
The unit price on the quote is the tip of the iceberg. In every WMS contract I've touched, the real pain comes from these four — skip them in negotiation and the signature becomes the trap:
Minimum commit. Many per-order contracts carry a floor: $5,000 a month minimum, paid whether or not your volume gets there. Three slow months means $15,000 for nothing. Ask whether the floor can settle quarterly, or roll unused volume forward.
Tiered overage. That "$0.12 per order" usually comes with fine print: "volume above 100,000 orders per month billed at $0.15." A 6,000-order-a-day warehouse runs about 156,000 orders a month over 26 working days — the extra 56,000 all bill at the higher tier. The real blended rate isn't $0.12, it's $0.135. Make the vendor lay out the full tier table, then run your own historical volume through it.
Implementation fee. $50,000-80,000 one-time is standard in this industry. But check what it covers: are integrations included? How many training days? Is the 3-month hypercare after go-live billed extra? I've seen an $80,000 implementation where not a single EDI interface was included — another $24,000 on top.
Interface fees. Each marketplace, each EDI connection, each carrier label integration: thousands to tens of thousands per year, per connection. A warehouse on Shopify, Amazon, and Walmart can easily add $20,000-30,000 a year in interface fees alone. Ask: how much to add one more channel?
Add the four together and I've watched a $120,000 quote land at $210,000 in year one.
Which business fits which model
No universally good model exists. There's only fit.
Steady-volume e-commerce operators fit per-order. 8,000 to 10,000 orders a day, low variance — per-order is pay for what you use, and the math stays clean. Volume growth usually means revenue growth, so cost stays a stable ratio.
Seasonal businesses fit per-user. Toys, holiday gifts, outdoor gear — the whole year hinges on Q4. Under per-order, the Q4 bill alone can be 60% of the annual total; under per-user, cost is flat all year. I knew a holiday-decor warehouse in Los Angeles whose Q4 volume ran 6x Q1: per-order would have cost $220,000 a year, per-user $140,000.
For 3PLs with multiple clients, mirror how you bill your customers. If you charge shippers per order, a per-order WMS passes cost straight through and reconciliation is trivial — one-to-one. If your shipper contracts are fixed fees per pallet or per slot, a per-order WMS eats your margin, and per-user fits better.
One more that's easy to miss: highly automated warehouses — few people, huge volume — should pick per-user without thinking. Twenty people doing the work of sixty, and per-order pricing punishes you for the efficiency gain. The most extreme case I saw: a heavily automated e-commerce site, 35 users supporting 20,000 orders a day — $168,000 a year per-user versus $620,000 per-order.
Three questions for the negotiating table
Whichever model you pick, asking these three out loud saves real money:
One: "Can the minimum commit go away — or settle on quarterly totals?" Vendors rarely kill it outright, but they often agree to quarterly true-ups, letting slow-month volume shift into peak months.
Two: "Is the user count named users or concurrent users?" This is the biggest trap in per-user pricing. Named users means one license per head, with add/remove paperwork on every hire and departure. Concurrent means simultaneous logins — 25 heads might need only 15 concurrent licenses, an instant 40% saving. Push for concurrent; if that fails, push for flexible license add/remove terms.
Three: "What's the three-year total cost of ownership, with implementation, interfaces, and escalation clauses included?" Plenty of contracts bake in 5-8% annual increases. Make the vendor put three years of numbers on the table — you'll find the cheapest year-one option is often the most expensive by year three.
My verdict: volatile volume goes per-user, stable volume goes per-order, and 3PLs follow their shipper billing model. But whichever you choose, multiply the quote by 1.5 — that's your real year-one number. The WMS salesperson isn't lying to you. He's just not telling you everything.
