Writing a Wholesale Allocation Policy That Holds Up in Apparel

It is Tuesday morning during pre-book delivery week. A buyer at one of your top five wholesale accounts emails to ask why their PO shipped at 78 percent fill when the rep promised 95. Customer service pulls the order, sees that two SKUs were short, and discovers that those same SKUs shipped in full to a smaller account that placed its PO three days later. The sales rep does not know why. The warehouse picked what the system told it to pick. The planner says the allocation ran overnight and there is nothing to do now. Somewhere between the PO and the pick ticket, a decision got made that nobody can defend.

What is a wholesale allocation policy in apparel, and why does it matter?

A wholesale allocation policy apparel teams can actually use is a written, system-enforced set of rules that decides which order receives which units when demand exceeds available inventory. It covers pre-book POs against production commitments, at-once orders against on-hand stock, replenishment against safety inventory, and the line between wholesale-committed pools and DTC ATS. It defines tiebreakers when two orders compete for the last cases of a SKU. It names who can override the rules, under what conditions, and where the override gets logged.

The policy matters because allocation is the single decision in wholesale operations where the gap between what you promise and what you ship gets settled. Get it wrong and the consequences land on your largest accounts first, because they are the ones whose POs sit at the top of the queue with the most lines, the most cases, and the most exposure to a partial shipment. Get it right and you can defend every fill rate variance with a paper trail.

In the 6 Breakpoints framework, this lives squarely at Breakpoint 4: order flow becomes harder to trust. When allocation is not a written, enforced policy, the order-to-cash chain breaks at the point where promises convert into picks.

Why does allocation quietly fail at $10M to $20M?

The objections I hear most often in evaluations are not really about allocation logic. They are about the symptoms allocation failure produces: chargebacks creeping toward 1 percent of wholesale revenue, fill rates that vary 15 points between accounts of similar size, sales reps who have lost faith in the system and now negotiate units over Slack. Behind every one of those symptoms is the same root cause. There is no policy. There are habits, a few overrides the planner remembers, and a spreadsheet that two people maintain.

The failure mode is consistent in the $10M to $20M band. Brands at this size are running pre-book and at-once and DTC at the same time, often with a 3PL doing the picking. Inventory truth is already strained, which is Breakpoint 3, so the allocation engine, if there is one, is operating on numbers that are 24 to 48 hours stale. The planner compensates by holding back arbitrary safety stock for the accounts they remember to protect. Everyone else gets whatever is left after the manual overrides settle. On a $15M brand running wholesale, DTC, and 3PL together, this kind of reconciliation absorbs 6 to 9 hours a week from one person who should be doing planning, not data plumbing.

What I see from prospects who have already shortlisted three vendors is that they describe the problem in vendor-shopping language, asking which platform has the best allocation engine. The better question is which platform forces them to write the policy down. The engine is downstream of the policy. Without the policy, the engine just automates whichever bad habit got encoded first.

What does a defensible allocation policy actually contain?

A policy that holds up in practice has six components. None of them are theoretical. Each one resolves a specific argument that will otherwise happen by email at 4pm on a Friday.

First, channel pools. The policy names how on-hand and on-order inventory is segmented across wholesale-committed, DTC, marketplace, and at-once at-risk. A pool is not a wish. It is a number that the order entry screen and the DTC site both read from. If wholesale pre-book has been committed against units in a specific PO from the factory, those units do not appear in DTC ATS until pre-book ships. This is the part most brands skip and then wonder why DTC oversold the new drop while wholesale fill dropped to 82 percent.

Second, prioritization tiers. Accounts are ranked by criteria the merchant team agrees with in advance. Tier criteria can include annual volume, payment history, chargeback rate, retailer compliance score, or strategic value (a flagship majors door, for example, will outrank a stronger volume account if the brand has decided distribution matters more than revenue this season). The point is not which criteria you pick. The point is that the criteria are written down and the system can sort against them.

Third, tiebreakers. When two accounts in the same tier compete for the same last cases, the policy names the order: PO receipt date, requested ship date, order completeness if shipped at current inventory, then account size. A four-line tiebreaker is enough to settle 95 percent of the arguments that would otherwise escalate.

Fourth, cancel and substitution rules. Every PO has a cancel date. The policy names which units the allocator can substitute (size up or down in the same colorway, alternate colorway in the same style) and which it absolutely cannot (no style substitutions, no color substitutions in core programs). It names the threshold below which a partial ship is held versus released. A common cutoff is 70 percent line fill on majors accounts, 60 percent on specialty.

Fifth, override authority. Two people, named by role, can override the allocation. Every override writes a record with the user, the timestamp, the original allocation, the new allocation, and the reason code. This is the single most important piece, because it converts allocation from a verbal argument into an auditable decision.

Sixth, the calendar. Allocation runs on a published schedule. Pre-book allocates against factory commitments on a defined date relative to the ship window. At-once allocates daily against available-to-promise. Replenishment allocates against safety stock thresholds. Sales reps know the schedule. Buyers know the schedule. Nobody has to ask whether their PO has been allocated yet.

How does DTC fit into a wholesale allocation policy?

Wholesale should not run through Shopify’s native flow, and DTC ATS should not include units that wholesale has already committed against. Those are two statements of the same principle: the channels share inventory but they do not share decision-making.

In practice, this means the DTC site reads from a channel-aware available-to-sell number, not from total on-hand. If 1,200 units of a SKU are on hand but 900 are committed against open wholesale POs and 100 are reserved for replenishment to a key account, DTC sees 200. The reason most brands oversell at peak (the 2 to 3 percent oversell rate is a back-of-envelope for a $15M brand running both channels through a 3PL) is that the DTC site is reading total on-hand and the wholesale commitments are invisible to it.

This is not a Shopify problem. It is a policy problem that Shopify cannot solve on its own because Shopify does not know what a wholesale commitment is. The allocation policy has to define what counts as committed, when the commitment is created (at PO entry, at allocation run, or at pick), and how DTC ATS is calculated against it. Once the policy is written, the integration is mechanical. Until then, the integration is a permanent argument.

What happens to chargebacks when the policy is missing?

If retailer chargebacks exceed 1 percent of wholesale revenue, the EDI integration is usually not the problem, and neither is the warehouse. The problem is upstream. Allocation decisions made manually, late in the cycle, produce ASN files that do not match what the buyer expected, short ships that breach the cancel window, and substitutions that violate retailer compliance manuals. Every one of those is a chargeback line item.

A written allocation policy reduces chargebacks because it forces the cancel window into the allocation logic, not into the rep’s email. If the policy says no partial ships below 70 percent on majors and the order will not hit 70 percent before cancel, the system flags it as a cancel candidate three weeks early, not three days late. The buyer can re-cut. The brand does not eat a chargeback for short ship and a chargeback for late ship on the same PO.

Lufema, a multi-entity wholesale operator, runs across multiple brands and 100-plus retailer accounts, and their inventory accuracy now sits around 99 percent, up from the 90 to 95 percent range they had before they consolidated onto a single operations system. The accuracy gain is partly inventory truth at Breakpoint 3 and partly allocation discipline at Breakpoint 4. They onboarded three new brands and over 100 retailer accounts without adding operations headcount, which is not possible if allocation is happening in spreadsheets and email.

When should the policy run, and how often should it change?

The policy itself should change once a season, not once a week. The allocation runs should be on a published cadence: pre-book allocation locks at a defined date before the ship window opens, at-once runs daily, replenishment runs twice a week against safety stock. The cadence is part of the policy, not a separate operations decision.

Where brands get this wrong is by treating every season as a fresh negotiation. The tier list gets revisited mid-season because a rep is fighting for an account. The cancel window gets relaxed because a buyer asked nicely. Three weeks later, nobody can remember what the rule actually is, and the planner is back to deciding by feel. The policy is supposed to be the thing that does not change inside the season. Change it between seasons, in writing, with the merchant team and the ops team in the same room.

What is the architectural fix?

The fix is not a better allocation algorithm. The fix is moving allocation out of the spreadsheet and the email thread and into a system where the policy is enforced, the overrides are logged, and the channel pools are honored across wholesale and DTC. This is the work of Breakpoint 4: making order flow trustworthy enough that everyone downstream (the warehouse, the buyer, the rep, the controller) can act on what the system says without verifying it manually.

For most $10M to $20M brands, this is also the moment where the cost of running on disconnected tools becomes visible. The planner spends 6 to 9 hours a week reconciling. The chargebacks climb. The DTC site oversells at peak. The 3PL ships what it is told and is blamed for decisions made (or not made) two systems upstream. Replacing the 3 to 5 tools and the spreadsheets with a system that enforces the policy is not a technology decision in the abstract. It is a decision to stop relitigating allocation every Tuesday.

What this means for an apparel operations team

The team that needs to act on this is not just operations. The merchant team owns the tier criteria. Sales owns the account list. Finance owns the chargeback exposure. Operations owns the enforcement. If any of those four are not in the room when the policy is written, the policy will not survive the first season.

Start by writing down what the current policy is, even if the current policy is whatever the planner decided last Tuesday. Then audit one week of allocation overrides and ask whether each one was defensible. The pattern that emerges is the policy you actually have. The policy you want is one step removed: same shape, but with the tiebreakers explicit, the override authority named, and the channel pools enforced in the system rather than in someone’s head.

The brands that get this right do not have better allocation engines than the brands that do not. They have written policies, enforced systems, and a calendar everyone in the building can name. That is the difference between Breakpoint 4 holding and Breakpoint 4 failing.