Most Amazon vendors file direct fulfillment under one of three headings: FBM for 1P, a bit of backup inventory, or a clever way to make chargebacks disappear. All three undersell it, and the last one gets people into trouble.
On an episode of MerchantSpring Marketplace Masters, Chris Khoo, founder of KhooCommerce and someone who has spent more time inside Amazon vendor operations than almost anyone, made the case that direct fulfillment (DF) is really an operations discipline. Run it well, and it protects your buy box and your peak-season revenue. Run it badly, and Amazon will quietly stop sourcing from you.
A cleaner way to think about it, and the term Amazon uses internally, is dual fulfillment. You keep your regular purchase-order flow into Amazon's warehouses, and you add a second stream that ships direct from your warehouse to the customer. The point is redundancy: bridging out-of-stock gaps so you always own the buy box and the shopper can always buy. Here is what it actually takes to do that well.
The dominant pitch for DF is chargeback reduction, often quoted as 3 to 5% of savings when you migrate off pure PO-based ordering. Khoo's view is blunt: if you switched an entire 1P operation to DF purely to dodge chargebacks, you would surface other business costs that wipe out the case. Fewer compliance headaches are a genuine upside, not the reason to do it.
The real reason is availability. When inbound placement times, appointment slots or mismatched purchase orders leave you unable to keep stock in front of customers, DF lets you smooth the demand curve yourself, especially through peaks. Khoo describes it flexing from a quiet 10 to 15% of volume to suddenly absorbing a surge: “this week DF just took off, and we did another hundred thousand dollars that we would have basically lost because Amazon didn't have it in stock, and we've got it.” That is the argument for DF, and it is a revenue argument, not a cost one.
Khoo's shorthand for a great DF candidate is bulky, high value, or awkward to warehouse. Large and medium domestic appliances (washing machines, microwaves), furniture, TVs, pool tables, garden sets: customers already expect a few days' delivery, the price tag is high enough to carry the cost, and you avoid the loss points of shipping goods into Amazon and back out again. Multi-part products are a second strong case. A bed frame that arrives as three boxes tends to get fragmented inside Amazon's fulfillment centers; fulfilling it directly gives you control. Seasonal and spike products are the third: during a UK heat wave with air-conditioning units sold out everywhere, the vendor who can ship direct from the warehouse captures demand that a purchase-order cycle would miss entirely.
The value threshold matters because DF adds handling cost. If picking and dispatching an order costs you around five pounds in labour, you will not want to run it on anything under roughly fifty pounds retail. A $100 microwave or a $300 to $500 dishwasher absorbs that handling comfortably; small, light items like accessories, notebooks, pens or footballs generally do not. For those, DF might be worth keeping open as a channel, but it is not where you should spend your operational energy.
The bigger question is not the product category. It is whether your operation can actually fulfil single orders to consumers. Vendor Central businesses are built around the wholesale rhythm: purchase-order cycles, ASNs, palletised shipments, credit terms, a few days of lead time to process a truckload. DF sits under that same financial model, but operationally it behaves like e-commerce. It is drop-ship processing, order by order, against a tight clock.
So the honest test is whether your warehouse, in-house or third-party, is willing and able to handle direct-to-consumer picks, and whether it can absorb a sudden peak (five hundred air conditioners tomorrow) without falling over. If you already run a smooth e-commerce or marketplace operation (Shopify, eBay, TikTok Shop), DF is conceptually just another one of those. If you do not, you will get killed on the performance metrics before the strategy ever pays off. Stress-test the warehouse before you enrol, not after.
Unlike the cyclical PO spikes that land Monday or Wednesday, DF orders arrive continuously. The moment a shopper clicks buy on an item you have won through DF, it lands on your side, and a well-run operation hits three checkpoints: it stays coherent to the customer, it hits Amazon's metrics, and it doesn't gum up internally.
In practice, that means a clean integration between your ERP and Amazon. Orders come in, you run a stock check (right product, is it available), and you send an acknowledgement back to Amazon. The SLA is within 24 hours, but the best operators do it as fast as possible. Automation then funnels each order to the warehouse as a picking instruction with box counts and labels pulled from Amazon. Rather than single picks, most brands batch: sweep the last few hours of orders (often one to three hundred at a time), group them by product code for a combined pick, then batch-print labels and stage everything freight-ready. The whole day is timed around carrier cutoffs. If UPS collects at 2 pm, the morning batch is picked and pulled well ahead, with a final sweep around 1 pm. What separates a smooth operation from a chaotic one is unglamorous: integration and connectivity, automation, and honest internal conversations about who does what when volume spikes.
DF has its own metrics, and the one that matters most is on-time shipping compliance: how closely your dispatch date matched Amazon's required ship-by, measured on a trailing average. The stakes are higher than on the retail side. DF behaves more like Seller Fulfilled Prime, so if you repeatedly miss, you are not just eating a chargeback. Amazon can decide you are not operationally cut out for it and simply stop sourcing DF from you. There are also charges for confirming an order and then cancelling it late, analogous to the down-confirmation or not-filled chargebacks on the retail side, which come down to picking accuracy and stock accuracy, both fixable through good integration.
One important reassurance: DF is not a 24/7 obligation. You set your warehouse's operating hours, weekends and public holidays included, and Amazon works expected ship dates around them. An order placed Thursday evening before a long weekend can carry a Monday ship date. Most orders ship the same or next business day, but the required date can sit further out, which changes how you plan the queue.
On carriers you choose between a vendor-owned carrier (VOC) and an Amazon-owned carrier (AOC), set at the account level (for a full run-down of these and other vendor acronyms, MerchantSpring keeps a retail and Amazon vendor acronym glossary). Most operations use Amazon-owned: they order a unit, you confirm the box count, they hand you a label, almost always UPS. Because Amazon owns the shipping, it pays and it carries the liability from the moment of pickup. If you own the carrier, you wrap the cost into your dispatch margin and hold liability longer. For most vendors, “just give me the label” is the pragmatic choice.
On integration, you have EDI, the SP-API direct fulfillment endpoints, or manual uploads. Functionally, Khoo sees little difference between EDI and API, and tends to favour EDI because it is wider-reaching, while some DF API endpoints are still maturing (Amazon SP-API vendor direct fulfillment docs). Starting manual and integrating once volume justifies it is pragmatic, with one caveat: leave it too long, get swamped, miss compliance, and you can damage your standing with Amazon. The feed people overlook is the inventory feed from your ERP up to Amazon. It publishes what you have available and drives the customer-facing availability view, so if it is wrong, shoppers can't buy at all.
Amazon often nudges certain categories toward DF because it does not want washing machines or flat-pack furniture clogging its warehouses. Exactly how much Amazon saves is opaque, so Khoo's advice is to treat any push (DF, WePay, direct import) the same way: ask why. Is it margin, a vendor manager's scorecard, or a genuine response to your out-of-stocks? Then aim for a win-win. If your product is a good candidate and you would still make decent money, look at it, but push back on price by framing your real added costs (integration, handling, packaging, downstream warehouse load) rather than simply conceding on Amazon's savings. It is still the wholesale price Amazon buys at, even though it is the retail price on the front end, and as Paul Sonneveld noted on the show, some rebates, co-op and deductions fall away under DF, which leaves room to negotiate.
The failure Khoo sees most is a ramp-up problem. A brand dips a toe in, runs one order a day through a third-party warehouse that is perfectly happy, and then Prime Day plus a seasonal spike drives three to four hundred orders in two days. Suddenly the warehouse is asking what was contractually agreed and who owns the extra manual work. Nothing was done wrong, exactly; the miss was communication and planning. The lesson: build for where you are today, but have a real view to scale.
At the other end, Khoo describes a glassware brand for whom direct shipping is ideal (fragile, high value, better off avoiding extra handling). They run three to four hundred orders a day through batch picking with customised labels that carry the picking instruction, so operators simply print, pick, pack, label and pull. No spreadsheet downloads, no cross-checking; it flows straight into NetSuite with item fulfilments updating automatically. DF is not half their Amazon business, more like 80-20 against their vendor side, yet at their scale that is still millions of dollars a month, run alongside their regular POs. The common thread across the best operations is the same three things: integration, culture, and operations. For more on how high-performing vendors approach this, MerchantSpring's guide on what high-performing vendors know about Amazon direct fulfillment is a strong companion to this piece.
DF only earns its place if you can see its true contribution next to your regular POs, net of the extra handling. MerchantSpring flows direct fulfillment purchase orders into its Vendor Profitability reporting, so your P&L reflects what DF is actually adding, and its order management lets you confirm shipments and pass tracking back to Amazon in one place. See it on your own vendor accounts with a walkthrough of your data.