DSD vs. Warehouse Distribution: What the Numbers Show
Consider a brand shipping the same product through two channels. Through UNFI, the headline margin is 25%; by the time slotting fees, promotional allowances, data fees, spoils charges, and freight markups have landed, the true cost stack runs closer to 50% of wholesale. Through a regional direct-store-delivery (DSD) distributor, the margin is 28%, and that number is close to real, because DSD distributors do not pile on the same ancillary fees.
The brand's COO looks at the two figures and asks why anything goes through UNFI at all. The answer is reach. The DSD distributor covers 40 stores in two metro areas; UNFI covers tens of thousands of retail doors nationally. The cost comparison is real. So is the constraint it ignores.
The cost stack is not a single number
Warehouse distribution through UNFI or KeHE begins with a cost-plus markup, typically around 10% on top of the brand's wholesale price, sometimes more, depending on volume tier and category. That is the visible number. The invisible one is the "inside income": a stack of fees the distributor collects on top of its margin.
Opener's analysis of a 50,000-case annual program at $24 wholesale itemizes the stack: $6.00 per case in distributor margin, $0.60 in free fills, $0.80 in slotting, $2.40 in promotional allowances and manufacturer chargebacks, $0.50 in co-op marketing, $0.30 in data and EDI fees, $0.40 in spoils, $0.90 in freight, and $0.70 in chargebacks and deductions. The true landed net: $11.40 per case, 52% of wholesale consumed by distribution.
DSD economics are simpler. The margin runs 25-35%, depending on category and route density. Free fills still apply; promotional allowances exist but are smaller. The data fees, EDI charges, and slotting structures that define warehouse distribution are largely absent. A brand running $500K through a regional DSD partner at a 30% margin pays $150K. The same $500K through a national distributor's full cost stack might net the brand $240K, or less.
The comparison is not quite apples to apples. Under DSD the brand absorbs costs the warehouse model pushes onto the distributor: merchandising labor, route logistics, delivery scheduling, shelf-condition monitoring. Those costs are real. But they sit visibly on the brand's own P&L, which makes them far easier to manage than a fee stack that arrives, months later, as deductions on a remittance.
DSD controls what warehouse distribution cannot
A DSD route driver stocks the shelf, rotates product, checks for out-of-stocks, and adjusts facings. A warehouse-distributed product arrives at the retailer's distribution center, is allocated to stores by the retailer's replenishment system, and sits on whatever shelf the store team gives it.
The difference shows up in the data. A GMA study of DSD economics, conducted by AMR Research and Clarkston Consulting, found that DSD products carried 2-4% lower out-of-stock rates than products moving through retailer DCs. In a category where the velocity number buyers watch in line reviews decides whether a SKU keeps its shelf space, a 2-4% out-of-stock improvement is no logistics footnote. It is the difference between the buyer seeing 2.1 units per store per week and 1.8, and 1.8 sits below the threshold that triggers a facing reduction in most natural-channel category reviews.
DSD also sidesteps compliance exposure. A shipment to a UNFI or KeHE DC is subject to appointment scheduling, ASN accuracy requirements, and fill-rate thresholds, the OTIF regime that generates chargebacks whenever the data describing a shipment fails to match the shipment itself. A DSD delivery to a retail store carries no ASN, no EDI 856, no OTIF penalties. The brand's compliance cost on that channel is zero.
Warehouse distribution scales; DSD does not
The same GMA study reported that DSD accounted for 24% of grocery unit sales but 52% of retail profits. Those profits accrued to firms with national DSD infrastructure (PepsiCo, Coca-Cola, Frito-Lay) companies running thousands of trucks in every market in the country.
A $15M specialty food brand runs no such fleet. A regional DSD operation covering 30-80 stores in two or three metro areas is manageable. A national one covering 500 stores demands trucks, route-optimization software, regional warehousing, and a labor force that a brand at this size cannot fund.
Hence the inflection point. A brand that starts at farmers markets, grows through regional DSD, and wins authorization at a national chain cannot serve that chain by truck. It enters warehouse distribution (UNFI, KeHE, or the retailer's own DC) and the cost stack, the compliance requirements, and the data infrastructure all change at once. The brand that ran its demand view off DSD route reports is suddenly reading shipment records in the ERP that tell a different story than retail scans.
The hybrid is where most brands land
Most specialty food brands at $8M-$25M run both models: DSD for independents, co-ops, and regional chains, where shelf condition and velocity matter more than cost efficiency; warehouse distribution for national and large regional chains, where reach and compliance are the price of entry.
The catch is that two distribution models produce two data streams. DSD revenue arrives as direct invoices. Warehouse revenue arrives as distributor settlements, complete with deductions that require their own reconciliation. Trade spend runs through both channels but is committed and measured differently in each. A brand that does not reconcile the two streams is managing its P&L from two versions of the truth, and the version that comes through the distributor is the harder one to read.
Map the distribution cost before the next line review
Lailara builds distribution cost reconciliation for brands running DSD and warehouse distribution in parallel, true cost per case by account, compliance exposure by channel, and velocity comparison across distribution models. The deliverable is a per-account margin view showing where each distribution dollar goes and which accounts are earning their cost stack. Book a 30-minute scoping call.