Amazon SFP Trial: What to Fix Before You Start

You can execute every shipment correctly during your Seller Fulfilled Prime trial and still fail. That actually happened to one seller we worked with. Their shipping execution was clean, their tracking was clean, the warehouse was getting the orders out, but their 1-day delivery promise performance was stuck at around 15%. And when we investigated, the warehouse wasn't the real problem. There were 4 different problems hiding upstream. First, their integration was sending a 2-day handling time to Amazon. Second, some of their ASINs had been assigned to an Amazon-created default Prime shipping template instead of the Seller-Fulfilled Prime template they intended to use. Third, some important SKUs didn't have inventory in the locations needed to show fast delivery promises. And fourth, a meaningful amount of their advertising traffic was arriving after cutoff when customers could no longer see the same fast delivery promise. So they did a lot of the actual shipping correctly, but Amazon was measuring a promise that their system wasn't consistently creating. That's the first thing you need to understand before starting a Seller Fulfilled Prime trial. The trial is not just a warehouse test. It is a test of the entire system that creates and then fulfills the Prime promise. And you should test that system before Amazon does. Amazon's current trial runs for 4 full performance weeks, Sunday through Saturday. Once it starts, you don't get a setup week. You don't want day 3 to be the day you discover the wrong shipping template, or day 7 to be the day you realize inventory never finished checking into your West Coast warehouse, or week 2 to be the first time somebody asks what happens when the normal carrier service cannot make the promised delivery date. The trial should validate your operating model. It should not be where you invent it. Start with the products. Do not put your entire catalog into the trial simply because those products are eligible. The best trial SKUs are boring. They have predictable demand, predictable handling, stable inventory, enough margin, and enough order volume that one or two exceptions don't destroy the entire sample. Here's why volume matters. Imagine you only ship 25 trial orders in a week. If one arrives late, you're at 96% on-time delivery. If 2 arrive late, you're at 92%. Amazon's current Seller Fulfilled Prime requirement is at least 93.5% on-time delivery. 2 carrier problems just changed the result dramatically. That doesn't necessarily mean your warehouse is bad. It means small samples are volatile. So, Don't choose a SKU just because you think it can generate barely enough trial orders. Choose products with enough demand to produce meaningful data and enough margin to survive the occasional ugly shipment. There's another complication. During the trial, your products do not have the Prime badge yet. So if you're depending on the Prime badge itself to create the order volume you need, you have a problem. You may need advertising, promotions, or other traffic already planned before the trial begins, and your marketing team needs to understand that traffic timing can also affect the delivery promises Amazon measures. Most fulfillment teams naturally start thinking at the order— pick it, pack it, buy the label, get it onto the truck— but Seller-Fulfilled Prime starts before the order exists. Amazon has to decide what delivery date to show the shopper, and that promise depends on things like the inventory Amazon believes is available, the shipping template, handling time, fulfillment location, cutoff time, and delivery region. So before your trial begins, test what Amazon is actually seeing, not what you think you configured. In that failed trial I mentioned earlier, the seller's integration was silently sending a 2-day handling time to Amazon, The warehouse could ship faster than that, but Amazon doesn't make a customer promise based on how fast your warehouse tells you it can operate. It makes the promise based on the data it receives. That's a huge difference. Then check every trial ASIN against its actual shipping template assignment. Don't just open your beautifully configured Prime template and say, yep, that looks right. Export or inspect which SKUs are actually assigned to it. The seller in our example discovered that Amazon had created a default Prime template and some ASINs had ended up there instead. Everything can look correct at the template level while the actual SKU is attached to the wrong template. And finally, act like a customer. Check representative trial products from different zip codes. Check them before cutoff. Check them after cutoff. Check the regions where you expect most of your customers to be. You're trying to answer a simple question: Is Amazon showing the delivery promise we designed this operation to produce? If the answer is no before the trial starts, that's good news because you still have time to fix it. Here is one of my favorite readiness tests. Ask yourself, if 50 Seller-Fulfilled Prime orders arrive tomorrow, could every fulfillment location immediately ship the orders assigned to it? Not did the truck arrive, not do we technically own the inventory, can the warehouse actually pick and ship it. Inventory that is sitting on a receiving dock is not ready. Inventory that hasn't been reconciled isn't ready. Inventory that exists physically but hasn't synchronized correctly into your order management system, warehouse management system, or Amazon is not considered ready. An inventory that exists in Pennsylvania when the order needs to ship economically from California doesn't help you very much either. The dangerous thing is turning on the Prime configuration because your network looks ready on a spreadsheet while the actual inventory is still moving through the system. Then orders begin arriving and the team starts rescuing them. Maybe you reroute manually, maybe you ship from the wrong warehouse, Maybe you upgrade to Air. Maybe you cancel something because the inventory wasn't actually available. Those are expensive ways to discover that you launched 3 days too early. If you fail the 50-order test, don't turn Seller Fulfilled Prime on yet. Next, pressure test the physical operation. Amazon's minimum Seller Fulfilled Prime cutoff settings require 2 PM or later local time for Monday through Friday. and 10:30 AM or later on weekends. But simply putting those hours into Seller Central doesn't make the operation ready. If you say you operate on Sunday, who is picking? Who is packing? Who fixes a label failure? And most importantly, does a carrier actually pick up the packages? A staffed warehouse with no carrier handoff can still leave your Prime packages sitting on the dock. So before the real trial, create problems on purpose. Run test orders close to cutoff. Create a label failure. Pretend the intended fulfillment location is out of stock. Test whether the order reroutes correctly. Ask what happens when the normal carrier service no longer meets the delivery promise. Ask what happens when the carrier misses a pickup. Ask who notices, and ask how quickly that person can do something about it. Seller Fulfilled Prime currently requires at least 93.5% on-time delivery, at least 99% valid tracking, and a cancellation rate no higher than 0.5%. Those numbers leave room for occasional exceptions. They do not leave much room for discovering basic workflow problems during the trial. The best time to find the broken label workflow is before it counts. There's one more mistake I see sellers make. They define success as passing the trial. That's not enough. You can pass Amazon's trial and still discover that Seller Fulfilled Prime is a terrible business model for that SKU. Here's an example of an SFP merchant we analyzed who operates from 5 fulfillment locations. Ground shipping averaged roughly $18 per shipment. Only about 2% of orders required an air upgrade, but those air shipments cost roughly $23 to $47 each. That program can work because those products have enough margin to absorb the exceptions. But imagine a product where every $40 overnight shipment erases the profit from several normal orders, and you don't have the luxury of shipping from 5 different locations. Your air shipments achieve Amazon's requirements, but you're bleeding margins doing it. So before the trial starts, define what success actually means. How much premium shipping are you willing to tolerate? What exception rate is acceptable? What does the SKU need to earn after fulfillment? At what point do you pause and investigate instead of continuing to buy your way out of problems? Your stop-loss rules should exist before emotions get involved. Because once you're halfway through a trial, it becomes very easy to say, let's just get through the next week. So before you start the Seller Fulfilled Prime trial, I would want 6 things to already be true. Number 1, your trial SKUs make economic and operational sense. Number 2, your Amazon handling time, shipping templates, and SKU assignments have been verified. Number 3, your inventory is physically received, synchronized, and available from the locations that are supposed to fulfill it. Number 4, your warehouse has already rehearsed cutoffs, weekends, carrier pickups, routing, labels, tracking, and exception recovery. Number 5, you know how you're going to generate enough trial volume even without the Prime badge. And number 6, you already know what financial and operational result would make you continue or stop. Stop. Then start the 4-week trial. At that point, you're not hoping the operating model works, you're verifying that it works under real Amazon traffic. And ideally, the trial should be boring. Orders route correctly. Inventory is available. The correct carrier service appears. Labels work. Tracking flows back to Amazon. Exceptions are caught before they become defects. The team watches the metrics every day. Nobody is improvising. That is what readiness looks like. If this is your first time here, I'm Manish Chowdhary, founder and CEO of Kahoot! We've spent years working with e-commerce brands on Seller Fulfilled Prime and distributed fulfillment, including helping sellers diagnose exactly these kinds of trial and operating problems. And if you're preparing for a Seller Fulfilled Prime trial, We've put the checks I covered here, plus the SKU economics, inventory, warehouse, carrier, and go or no-go questions into an SFP trial readiness checklist. You can find the link below. But whether you use our checklist or build your own, remember this: don't use the trial to discover whether you're ready for SFP. Test the operation first, fix the weak points, then let the trial Confirm what you already know.