Five signs your EDI setup cannot scale past $15M in wholesale

It is a Tuesday morning at a $14M contemporary brand and the wholesale ops lead is on her third coffee. Nordstrom just kicked back an 856 because the ASN arrived four hours after the trailer left the 3PL. Dillard’s is asking for an 846 refresh they were supposed to get every 24 hours. The EDI broker’s portal shows two 850s stuck in a validation queue from Friday. The 3PL is emailing pick confirmations as PDFs because their WMS cannot push a clean 856 back to the broker on time. Nothing here is broken in a way anyone can point to. Everything is broken in a way that costs money every week.

Why does an EDI setup cannot scale apparel wholesale past $15M?

The short answer is that most EDI setups under $15M are held together by a broker plus a person. The broker translates the files. The person patches everything the broker cannot see. That arrangement works when you have four or five retailer accounts, one warehouse, and predictable ship windows. It stops working when you have twenty accounts, a 3PL, drop cycles, and retailers who each interpret the X12 standard slightly differently.

An EDI setup that cannot scale apparel wholesale past $15M is one where the electronic data interchange layer is disconnected from the systems that actually hold inventory truth, order state, and warehouse execution. Files move. State does not. The broker sends an 856, but the inventory system does not know the shipment left. The 850 comes in, but the allocation engine, if there even is one, does not reserve against it. Every retailer transaction becomes a small reconciliation project, and reconciliation projects compound.

This lives inside breakpoint four of the order flow, where order state stops being trustworthy. EDI is where BP4 becomes visible to the retailer, which is why the chargebacks show up before the internal dashboards do.

What I keep seeing when I look at wholesale ops in the $10M to $20M band

When we started Uphance, the recurring pattern in the brands I talked to was not that they lacked tools. They had tools. They had a broker, they had Shopify for DTC, they had a 3PL with a portal, and they had QuickBooks or Xero on the finance side. What they lacked was one place where an order actually lived from PO receipt through pick, pack, ASN, and invoice, with the EDI transactions treated as part of the order’s life instead of as a parallel file exchange happening in a different building.

That gap is what customers are actually buying when they replace their setup. They are not buying EDI. They are buying the elimination of the human whose job is to make EDI and inventory agree at the end of every day. On a $15M brand with wholesale, DTC, and 3PL, that reconciliation work runs six to nine hours a week and effectively consumes one FTE across ops, customer service, and finance combined. The chargebacks are a separate line item on top of that.

So when a wholesale ops lead asks whether their EDI setup can carry them past $15M, the honest test is not whether the broker is good. The test is whether the five failure modes below are already showing up.

Sign one: chargebacks are creeping past 1 percent of wholesale revenue

Retailer chargebacks are the earliest, cleanest signal that EDI and warehouse execution are out of sync. Every major department store publishes a compliance matrix. Late ASN, wrong ASN format, carton label mismatch, PO number mismatch, ship window miss, routing guide violation. Each has a fee. Individually they are small. In aggregate they scale with your wholesale volume, which is exactly the wrong shape.

If your chargebacks are running above 1 percent of wholesale revenue, the EDI integration is the problem, not the warehouse. The warehouse is executing. The system telling the warehouse what to execute, and telling the retailer what got executed, is failing at the handoff. On a $15M brand doing $9M wholesale, 1 percent is $90K a year of pure margin drag. At $25M wholesale it is $250K. That is a full ops hire that exists only because your 856 is three hours late.

The architectural fix is not a better broker. It is order state and shipment state living in the same system, so the ASN goes out the moment the pick is confirmed, not after someone reconciles two portals.

Sign two: ASNs depend on a person watching a 3PL portal

This is the failure mode I see most often between $10M and $20M. The 3PL confirms shipments through their own WMS. Someone at the brand logs into the 3PL portal, exports a shipment file, and forwards it to the EDI broker to build the 856. The broker sends it. The retailer receives it, sometimes on time, sometimes six hours after the truck showed up at the DC.

If this describes your flow, EDI is not integrated. It is being retyped. Every additional retailer account, every additional drop, every additional 3PL location multiplies the manual load linearly. There is no version of this that scales, and it is the first thing that snaps when volume steps up in Q3 or during a drop.

Lufema, one of our multi-entity wholesale customers, hit this wall while carrying multiple brands across a shared distribution operation. What moved their inventory accuracy from the 90 to 95 percent range up to around 99 percent was not a better 3PL. It was that pick confirmations, inventory movements, and outbound documents all lived on the same data. Onboarding three new brands and over a hundred retailer accounts without adding ops headcount was only possible because the ASN was no longer a manual step.

Sign three: the retailer sends 852 and 846 files that nobody reads

Most mid-market brands are set up to receive 850 purchase orders and send 856 ASNs and 810 invoices. That is table stakes. The transactions that separate a scalable setup from a stuck one are the ones going the other direction that most brands ignore.

The 852 is a product activity report. Weekly sell-through and on-hand at the retailer’s DCs and stores. The 846 is an inventory advice file, often used by retailers to request or confirm available inventory. If your setup receives these and does nothing with them, you are running your wholesale replenishment blind. Your buyer is guessing at reorders. Your allocation team cannot see where the sell-through is happening. Your OTB conversations are relying on the retailer’s buyer being generous with information.

A scalable EDI setup ingests 852s automatically and turns them into a sell-through view your merchandising team can act on weekly. Run OTB weekly during selling season, monthly is too slow, and the 852 is the input that makes weekly OTB possible without begging for a spreadsheet from the retailer.

Sign four: order acknowledgments and change requests are handled by email

The 855 is a purchase order acknowledgment. The 860 is a PO change. These are the transactions where the retailer expects a machine-to-machine response inside their SLA, and where mid-market brands almost always fall back to email because their broker charges per transaction type or their internal system cannot generate a clean 855 from the order record.

The symptom is subtle. It looks like a customer service rep spending an hour a day replying to buyers with confirmations, split-ship approvals, and partial cancels. The cost is that the retailer’s replenishment planner treats your brand as slower than the brands responding through EDI, and your fill rate score, which they do track, drifts down. A lower fill rate score means smaller subsequent POs. That is not a chargeback. That is revenue you never see.

A setup that scales generates 855s from the order the moment it lands, sends 865s for cancels, and handles 860 changes without a human retyping quantities. This is possible only when the order record and the EDI transaction record are the same record.

Sign five: inventory commitments do not reserve against open wholesale POs

This is the deepest failure mode, and it is the one that ties EDI back to inventory truth across channels. When an 850 comes in, does your available-to-sell drop immediately across every channel, including your Shopify DTC storefront? For most $10M to $20M brands running EDI through a broker plus Shopify plus a 3PL portal, the answer is no. The PO sits in the broker’s queue, or it lands in an order management layer that does not talk to the DTC storefront’s inventory feed.

The result is oversell. On a $15M brand at peak, the oversell rate runs 2 to 3 percent, and it is almost always concentrated in the SKUs that were promised to a wholesale account and simultaneously sold through DTC because the DTC channel never saw the reservation. Every one of those is a customer service call, a refund, a canceled wholesale line, or a chargeback for a short ship. All of it is avoidable.

Wholesale should not run through Shopify’s native flow, and DTC inventory should not be a separate ledger from wholesale inventory. Channel-aware ATS, where an incoming 850 immediately reserves against a wholesale-committed pool that is visible to but not sellable by the DTC storefront, is the architectural answer. It is also the point at which brands stop patching their EDI setup and start replacing it.

What the fix actually looks like

Every one of the five signs above is fixable in isolation. You can hire a better broker. You can build a middleware layer that polls the 3PL portal every fifteen minutes. You can pay per transaction for 855 and 860 support. You can build a reservation flag inside a spreadsheet the DTC team consults before publishing inventory. Brands do all of this, and it works for a quarter or two, until the next retailer onboards or the next drop lands or the next 3PL location opens.

The structural answer is that EDI, order state, inventory, and warehouse execution have to live in one connected system. Not integrated. Connected. The 850 lands as an order in the same database that holds the wholesale-committed inventory pool. The pick at the 3PL updates the same order record. The 856 generates from that record automatically inside the retailer’s ship window. The 810 posts from the same shipment event. The 852 comes back in and updates the sell-through view your buyer sees during Monday’s OTB meeting.

This is what the apparel operations platform layer actually solves, and it is why the $10M to $20M band is the predictable breakpoint zone. Below $10M, the volume does not force the issue. Above $20M, the cost of not fixing it becomes obvious in the P&L. In the middle, brands spend eighteen months in the failure modes above before deciding.

The wholesale scaling test

The question is not whether your EDI works today. It works, or you would already be replacing it. The question is whether it can absorb the next ten retailer accounts, the next 3PL location, and the next drop cycle without adding headcount or chargebacks. If any two of the five signs above are already visible, it cannot. If four or five are visible, the reconciliation cost is already funding a role that should not exist, and the fill rate scores at your top accounts are drifting in a direction you will not see for another two quarters.

The brands that clear $15M cleanly are the ones that stop treating EDI as a file exchange problem and start treating it as an order flow problem. That reframing is the one that changes what you buy.