By Bernardo Campelo, Forbes Business Council E-Commerce Leader, Amazon SPN Certified provider, Amazon SP-API authorized partner, and Founder of PrepVia.
A seller messages me most weeks with some version of the same plan. Their current 3PL missed a deadline, or the invoice came back with a line item nobody explained, and they are ready to move the entire catalog to a new facility on the strength of one good sales call. I ask them the same question every time. Have you sent a test shipment yet, and if so, what did you actually measure.
Most of the time the answer is no, or the answer is yes but the test proved nothing, because nobody wrote down what a passing test looked like before the truck left the dock. A forum thread on switching prep centers went around in July 2026, and the most upvoted reply in the whole thread was one sentence: send a batch of two hundred to three hundred units before you move anything else. That advice is correct. It is also almost useless without a design behind it.
A pilot shipment without a plan is just a smaller version of the same gamble. It tells you the new facility can receive a box. It does not tell you whether the facility can run your business.
The 60-second version
A pilot shipment only proves something if you design it before you ship it. Pick a SKU that resembles your real catalog, not the easiest carton in the warehouse. Write down what arrival, received, and shipped mean in hours before the truck leaves, and record six numbers against those definitions: dock-to-received time, received-to-shipped time, count accuracy, label quality, and the variance between the quote and the invoice. Two hundred to three hundred units proves competence at a small, quiet volume. It does not prove peak capacity, it does not prove how the facility handles an error because a clean run generates no error to watch, and it does not prove how the same process behaves at twenty times the volume. Treat the pilot as evidence for one decision only: whether this facility earns a second, larger shipment.
The Advice Is Right. Most Sellers Still Run It Wrong
The two hundred to three hundred unit number keeps showing up because it sits in a useful range. It is large enough to force a facility through its real receiving process, not a favor for a prospect. It is small enough that a mistake costs you a bad week, not a bad quarter. Sellers who already run a working relationship with one prep center and are now vetting a second one gravitate to this number instinctively, and the instinct is sound.
Where the plan breaks down is at the point of measurement. A seller ships the pilot lot, the units arrive at Amazon a few days later without incident, and the seller reads that as a passing grade. But arriving without incident is the easiest bar in the entire test to clear. A facility that has never seen your SKU can often get two hundred units out the door cleanly through sheer available slack, the same way a restaurant can handle a quiet Tuesday without proving anything about a Saturday rush.
This is a different exercise from a full switch. A full migration, the kind covered in how to switch 3PLs without losing a quarter, is about sequencing an overlap between two facilities so your selling inventory never sits exposed. The pilot comes earlier. Its only job is to generate evidence, before you write a single purchase order pointed at the new address.
Pick the SKU That Can Actually Expose a Problem
The single fastest way to waste a pilot is to send the easiest carton in your warehouse. A single unit configuration, no batteries, no fragile packaging, no multi-pack assembly, generic rectangular box. That SKU will sail through almost any receiving dock in the country and tell you nothing you did not already know.
A pilot earns its name when the SKU resembles the catalog you actually run. If a meaningful share of your volume is multi-unit case packs, send a multi-unit case pack. If you carry anything that needs poly bagging, bundling, or a kit assembly step, include one of those SKUs in the two hundred to three hundred units rather than saving it for later. If your catalog includes an item with a printed barcode instead of a clean FNSKU sticker, or packaging that tends to scan poorly, that is the unit you most need in the test lot, not the one you are tempted to leave out.
Choose one SKU that is genuinely representative and, if volume allows, a second that represents your hardest edge case. Two carefully chosen SKUs across two to three hundred units will teach you more than three hundred units of your easiest item ever could.
Set the Clock Before the Truck Leaves Your Dock
Every dispute over a pilot result traces back to the same root cause: nobody defined the words before the numbers came in. Arrival, received, and shipped sound self-explanatory until two people disagree about which one applies to a pallet sitting on a dock overnight.
Define three timestamps in writing, over email, before the pilot ships. Arrival is the moment the carrier delivers the freight to the facility, confirmed by a signed bill of lading or appointment log. Received is the moment the facility logs your units as counted and on hand in its system, not the moment the truck backs into the bay. Shipped is the moment the outbound carrier takes physical custody of the finished, labeled units bound for Amazon. Ask the facility to send you each timestamp as it happens rather than reconstructing it later from memory or from an invoice.
This single step separates a pilot that produces evidence from a pilot that produces a feeling. Two facilities can both claim a fast turnaround while measuring from entirely different starting points, and you will never know which one is faster until you force both onto the same clock.
The Six Numbers a Pilot Should Produce
A pilot shipment is not a pass or fail event. It is a small dataset. Collect these six numbers against the timestamps you defined in advance, and you will have something you can actually compare against your current provider or against a second candidate facility.
| Metric | What it tells you | Where the evidence comes from |
|---|---|---|
| Dock-to-received time | How long after arrival your units become a trustworthy, counted number | Signed bill of lading against the system receiving timestamp |
| Received-to-shipped time | How the facility actually performs once it is done counting, separate from how busy the dock was that day | Receiving timestamp against carrier pickup manifest |
| Count accuracy | Whether the declared quantity and the counted quantity match, unit for unit | Declared vs received line on the receiving report |
| Label quality and scan rate | Whether the facility own FNSKU printing and placement holds up under an Amazon scan, not just a visual check | Print quality and mis-scan rate on the shipment after Amazon receiving |
| Invoice variance | Whether the bill matches the quote you were given before you shipped anything | Line-by-line comparison of quote to final invoice |
| Total elapsed hours to Amazon receipt | The full clock a buyer or a finance team actually cares about | Arrival timestamp against Amazon shipment status |
None of these numbers means much in isolation. Together, across one pilot lot, they form a baseline you can hold every future shipment against, and a baseline you can compare directly to whatever your current facility is already producing.
Reading the Invoice Against the Quote It Sent You
The invoice is where a pilot quietly tells you the most, and where most sellers read the least carefully. A quote gathered during vetting almost always describes a best case: standard SKU, no exceptions, no rush handling. The invoice describes what actually happened to your specific units.
Pull the original quote and the final invoice side by side, line by line. Check the per-unit prep fee first, since that is the number every facility leads with and the easiest one to hold constant. Then check for anything that did not appear on the quote at all: a receiving fee tied to carton count rather than unit count, a storage charge that started accruing sooner than you expected, a labeling fee applied to units that already carried a compliant barcode. A facility worth a larger shipment can explain every line on the invoice against a term already written into its agreement or rate sheet. A facility that cannot is telling you how the next invoice, on ten times the volume, is going to read.
Compare the total against the current published rate for the same service, not just against what the sales rep quoted you verbally. A quote that undercuts a facility own public pricing by a wide margin on a two hundred unit pilot is rarely the number that survives your first full-size shipment.
24 to 36h prep. 35-hour end-to-end guarantee or the prep is free. Net-30 terms. From 50 units to full truckloads.
What a 300-Unit Pilot Can Never Prove
A pilot lot is a small, quiet sample, and a small, quiet sample cannot answer every question you actually need answered before you move real volume. Knowing what the test does not cover matters as much as reading the six numbers it does produce.
Peak capacity is the first gap. Two hundred to three hundred units moving through a facility on an ordinary week tells you almost nothing about how that same facility performs during the seven days before a major deadline, when every client shipment competes for the same dock doors and the same labor hours. A facility can run a flawless pilot in March and still buckle in a Q4 queue, because the pilot never tested the queue at all.
Error handling is the second gap, and it is the one sellers miss most often. A pilot that goes perfectly gives you zero data on how the facility behaves when something goes wrong, because nothing went wrong. The real question is not whether errors happen. Errors happen everywhere, eventually. The question is what the facility does the day a carton arrives damaged, a count comes up short, or a label prints wrong: who notifies you, how fast, and with what evidence. Watch how a facility talks about the handful of things that tend to go wrong during vetting, because a pilot alone will rarely show you.
Behavior at scale is the third gap. Process does not always move in a straight line as volume grows. A receiving team that comfortably counts three hundred units by hand may not comfortably count six thousand the same way, and a facility that never says so during a clean pilot is not being dishonest, it is simply describing a volume it has not yet been asked to run for you.
| What the pilot proves | What the pilot does not prove |
|---|---|
| The facility can receive, count, and ship your SKU correctly at a small volume | How the facility performs during a peak-week queue |
| The timestamps and evidence you asked for actually get delivered | How the facility handles a shipment that arrives with a real discrepancy |
| The quoted price matches the invoiced price at this volume | Whether pricing and process both hold at ten or twenty times this volume |
Turning Six Numbers Into a Go or No-Go Decision
Once the pilot lot is closed out at Amazon, sit down with the six numbers and your current provider own recent numbers on the same metrics, if you have them. A facility that beats your current provider on every measured number after one clean pilot has earned a second, larger test, not your entire catalog. A facility that matches or loses on more than one number has not earned anything yet, no matter how the sales conversation went.
The honest next step after a strong pilot is rarely a full migration. It is a second shipment, larger than the first, ideally timed to land during a busier week so you begin to see the facility under real conditions rather than the quietest ones. Save the full sequencing playbook, including how to run two facilities in parallel without a gap in your sellable inventory, for when the second test also holds.
- Choose a SKU that reflects your real catalog complexity, not the easiest carton on your shelf.
- Get arrival, received, and shipped defined in writing, with each timestamp promised to you as it happens.
- Confirm in advance how the facility will report count accuracy, down to the unit, not just the carton.
- Ask what evidence you will receive on label quality, including any mis-scans Amazon reports after receiving.
- Save the original quote before the pilot ships, so you have something exact to hold the invoice against.
- Decide your pass and fail thresholds for all six numbers before you see a single result, not after.
- Plan a second, larger pilot in advance rather than treating one clean run as a final answer.
Frequently Asked Questions
How many units should a prep center pilot shipment include?
Two hundred to three hundred units is the range most experienced sellers converge on, and it is large enough to force a facility through a real receiving and prep cycle rather than a favor for a prospective client. Fewer units risk a sample so small that a single miscount or a single lost carton distorts the entire result. A pilot much larger than three hundred units starts to carry real financial risk before you have any evidence the new facility deserves it.
What SKU should I send for a test shipment to a new 3PL?
Send the SKU that most resembles the bulk of your actual catalog, not the simplest carton in your warehouse. If a meaningful share of your volume involves multi-unit packs, poly bagging, or a printed barcode instead of a clean FNSKU sticker, include that complexity in the pilot rather than saving it for a larger shipment later. A pilot built entirely around your easiest item will pass cleanly and tell you almost nothing about how the facility handles the rest of your inventory.
What is the difference between a pilot shipment and switching 3PLs?
A pilot shipment is a small, defined test of two hundred to three hundred units used to generate evidence before you commit any real volume to a new facility. Switching 3PLs is the much larger sequencing decision that follows a successful pilot, involving a planned overlap between two facilities so your active inventory is never exposed to a gap. Treat the pilot as the evidence-gathering step and the full switch as a separate decision made only after that evidence holds up.
Does a clean pilot shipment prove a prep center can handle Q4 volume?
No. A pilot lot moving through a facility during an ordinary week says very little about how the same facility performs during a peak-week queue, when every client shipment competes for the same dock doors and the same labor hours. A facility can run a flawless two hundred unit pilot in a quiet month and still struggle when volume from its entire client base spikes at once. Ask directly how the facility staffs and prioritizes during its busiest weeks, since the pilot itself will not answer that question.
How do I compare a 3PL quote to the actual invoice after a pilot?
Pull the original quote and the final invoice and compare them line by line rather than checking only the total. Look first at the per-unit prep fee, then check for any charge that did not appear on the original quote at all, such as a receiving fee tied to carton count, an early storage charge, or a labeling fee applied to units that did not need it. A facility that can explain every invoice line against a term already written into its rate sheet or agreement is a stronger signal than a quote that looked attractive during vetting.
What does a pilot shipment not prove about a new prep center?
A pilot proves competence at a small, quiet volume and little else. It does not prove peak capacity, since a light week of freight tells you nothing about a queue during a major deadline. It does not prove how the facility handles an error, because a clean pilot generates no error to observe. It does not prove that the process holds at ten or twenty times the volume, since counting and staffing decisions that work at three hundred units do not always scale in a straight line to three thousand or six thousand.
What should I do after a successful pilot shipment before moving my full inventory?
Run a second, larger pilot before committing your full catalog, and time it if possible to land during a busier week so you see the facility under closer to real conditions. Compare the results against the same six numbers from the first pilot: dock-to-received time, received-to-shipped time, count accuracy, label quality, invoice variance, and total elapsed hours to Amazon receipt. Only after a second test holds up does it make sense to plan the full migration sequence and the inventory overlap that a real switch requires.
Final Take
The advice to send a test shipment before you move your business is correct, and it has been correct long before it turned into a popular forum reply. Where sellers lose the value of that advice is in treating the pilot as a pass or fail moment instead of a data collection exercise designed in advance. Two hundred to three hundred units, a representative SKU, three defined timestamps, and six recorded numbers turn a gamble into a comparison you can actually defend.
Just as important is knowing where the pilot stops. It will not tell you how a facility behaves in a peak-week queue, how it handles a real discrepancy, or whether its process holds at twenty times the volume you tested. Those questions belong to a second, larger pilot and to the vetting conversation itself, not to a single clean shipment.
If your current provider is already failing you on one or more of these six numbers, that is useful information too. A pilot shipment run against a new facility often becomes the clearest evidence of what your existing relationship has quietly stopped delivering.
Design the test before you ship it. The two hundred to three hundred units are cheap. The decision they inform is not.
See how PrepVia receives, times, and documents every shipment →
PrepVia is Amazon SPN Certified, prep window 24-36 hours, Net-30 available.
- How to Switch 3PLs Without Losing a Quarter: the full migration sequence once a pilot has already earned your trust
- Prep Center Agreement Checklist: the clauses that belong in writing before your pilot lot, or any lot, ships
- Prep Center Red Flags: the vetting signals a clean pilot will not show you on its own
