Most brands pick FBA because it is what everyone else does, then find out six months later, when low-inventory-level surcharges show up and half the catalogue's margin has quietly disappeared. Choosing between FBA and FBM is not a philosophical question. It is a subtraction, and you do it product by product.
What each model actually covers
FBA (Fulfilment by Amazon) means you send your stock into Amazon's fulfilment centres and they store, pick, pack, ship and handle customer service and returns for that order. You lose control of that inventory the moment it goes through the warehouse door.
FBM (Fulfilment by Merchant, you ship it yourself) means the stock sits in your warehouse or your 3PL's, and you handle everything: picking, shipping, tracking, incidents and returns. Amazon only charges you the category referral fee.
The difference that matters is not who moves the box. It is that with FBA you are buying speed and Prime eligibility in exchange for a cost structure you do not control and that Amazon revises every year.
The real cost of FBA sits outside the fulfilment fee
Almost everybody models FBA using the per-unit fulfilment fee and calls it done. That fee is the visible part. Underneath it there are at least five more layers:
- Monthly storage. Charged by cubic metre, and it rises sharply in the final quarter of the year. If your product is bulky and slow-moving, storage can eat more margin than shipping does.
- Aged inventory surcharge. Amazon penalises stock that has sat in the network for a long time, in bands that get worse the longer it stays. A batch that will not sell stops being a parked asset and becomes a recurring expense.
- Low-inventory-level and overstock surcharges. Amazon compares your available stock with your recent sales. Hold far more than you sell and you pay more. Hold too little and you go out of stock and lose position.
- Prep and labelling. If your product needs a polybag, tape, a label or bubble wrap and you do not send it ready, Amazon does it and bills you per unit.
- Returns. In many categories Amazon keeps part of the referral fee when it processes a return, and the returned unit may well end up graded as unsellable.
- Removal or disposal. Pulling out inventory that will not sell also costs money per unit.
Rule of thumb: if the fulfilment fee is the only FBA cost in your spreadsheet, your calculated margin is inflated.
The exact figures for all of this change often and vary by marketplace and by season. Always confirm them in Seller Central before you commit to a calculation. Do not trust a table you read on a blog, this one included.
When FBM wins outright
FBM is not the fallback for small brands. There are product profiles where it is simply the right answer, however big you are:
- Bulky or heavy products. Amazon's cost per cubic metre is high, and above a certain volume your 3PL will almost always beat it.
- Slow movers. If a line sells three units a month, storage and aged inventory surcharges swallow the sale.
- Fragile items or anything needing careful handling. In FBA it is handled by people who do not know your product. Breakages and damage-related returns go up.
- Products with expiry dates or batch control. Managing dates, FEFO (first expired, first out) and recalls inside FBA is awkward and expensive.
- Wide variant catalogues with uneven demand. Spreading FBA stock across forty sizes and colours forces you to tie up a lot of inventory just to avoid stockouts.
- Bundles and personalisation. If the order is assembled to order, FBA is not a realistic option.
The usual counterargument is "without FBA I don't sell". That is only half true. Without FBA you sell less if you do not make up for the delivery speed. With short dispatch times, a competitive carrier rate and a low order defect rate, a well-run FBM competes.
Buy Box and Prime: what really moves the needle
The Buy Box (the buy panel that takes the vast majority of sales on a listing with several sellers) is not won on price alone. Amazon weighs landed price including delivery, promised delivery speed, fulfilment reliability, cancellation and late shipment rates, and customer service metrics.
FBA starts ahead because Amazon trusts its own operation: it takes the delivery promise as given and does not penalise you for logistics incidents. Under FBM those metrics are yours to hold up. An FBM with valid tracking, next-day or two-day delivery and no late shipments can fight for the Buy Box. An FBM with five working days and no tracking cannot.
Prime is a separate matter. With FBA it is automatic. With FBM it does not exist unless you join Seller Fulfilled Prime. And losing the Prime badge hurts most in impulse-buy, low-ticket categories, where shoppers filter by Prime before they look at anything else.
Seller Fulfilled Prime: the middle road
SFP (Seller Fulfilled Prime) lets you carry the Prime badge while shipping from your own warehouse. In return Amazon demands a high, sustained operational standard: passing a trial period, using approved carriers with tracking, collecting at weekends, hitting the Prime delivery windows for each region and holding very tight on-time shipment and cancellation thresholds. The exact requirements and delivery windows by country get revised regularly, so check them in Seller Central before you decide anything.
SFP makes sense when you already have a mature logistics operation, enough volume to justify the effort, and a product where FBA destroys your margin but you still need Prime to compete. It makes no sense if your warehouse is not reliable day in, day out. Being removed from the programme for missing metrics hurts more than never having joined.
| FBA | FBM | SFP | |
|---|---|---|---|
| Cost structure | Per-unit fee + storage + surcharges + returns | Your warehouse cost + negotiated carrier rate | Your warehouse cost + express carrier rate (dearer than FBM) |
| Cost predictability | Low: Amazon revises fees and surcharges | High: you control it | Medium: exposed to express carrier pricing |
| Operational control | Minimal | Total | Total, within Amazon's rules |
| Delivery speed | Fastest by default | Whatever you can achieve | Prime standard, mandatory |
| Effect on Buy Box | Head start | Neutral or negative unless the operation is excellent | Comparable to FBA |
| Prime badge | Yes | No | Yes |
| Choose it when | Small, light, fast-moving product with comfortable margin | Bulky, heavy, fragile, dated or slow-moving product | You need Prime, FBA leaves you no margin and your logistics are reliable |
Run the maths per SKU, not per catalogue
The most expensive mistake is picking one model for the whole brand. Inside the same catalogue you will have lines where FBA is obvious sitting next to lines where it wipes out the margin.
Build a sheet with one row per SKU and these columns: selling price, category referral fee, cost of goods, estimated FBA cost (fee plus allocated storage plus expected surcharges), real FBM cost (pick, pack, box, carriage), that line's return rate, and units sold in the last 90 days. From there you get contribution margin per unit under both scenarios, and you multiply it by actual turnover. A SKU that earns 40p more under FBM but sells 2,000 units a month matters more than one that earns three pounds more and sells twelve.
Two criteria that help you sort the list:
- Turnover. Divide units sold in 90 days by units in stock. Below a certain threshold, FBA stops paying for itself once storage and surcharges are in.
- Value density. Selling price per litre of volume. The lower the number, the worse the product fits FBA.
As an industry reference point, across the accounts we manage it is common to find that somewhere between 15% and 30% of the lines in a broad catalogue are sitting in FBA losing money or making pennies. It is rarely negligence. It is that nobody has redone the maths since the catalogue went live.
A mixed model is the normal outcome
When the numbers are done properly, the answer is almost never "all FBA" or "all FBM". It is a split:
- FBA for the small, light bestsellers, where speed and the Prime badge convert.
- FBM for bulky items, the long tail and seasonal lines out of season.
- SFP, where it applies, for lines that need Prime but cannot absorb FBA's cost.
The mixed model has a management price: two inventory flows, two ways of reading available stock and more discipline in forecasting. In exchange, you stop using your bestsellers to subsidise the lines that lose money in FBA.
It also gives you a safety net. If a line runs out of FBA stock, having it live in FBM stops the listing going dark and losing sales history, which is the part that really takes time to rebuild.
What to do this week
- Pull the last 90 days of profitability by ASIN and cross it with your FBA storage and surcharge reports. Sort by absolute contribution margin, lowest first.
- Take the worst twenty lines and work out their value density (price divided by litres of volume). Flag everything below your catalogue's median.
- Ask your 3PL for a fixed pick, pack and 24-48 hour carriage rate for those lines, and compare it against the full FBA cost, not just the fulfilment fee.
- Remove from FBA any inventory that has sat longest and has no sales plan this quarter. The removal cost is nearly always lower than paying surcharges indefinitely.
- Before you consider SFP, measure your real on-time shipment rate and cancellation rate on FBM for four weeks. If they are not close to perfect, fix that before you apply for anything.