The standard advice is that FBA wins. It gets you Prime eligibility, it converts better, and it takes fulfillment off your desk. For most sellers, most of the time, that is correct.
But “most of the time” is doing a lot of work in that sentence. There is a specific and identifiable set of SKUs where FBM produces more net profit per unit and per dollar of working capital, and sellers who fulfill everything through FBA on principle are leaving money on those SKUs indefinitely.
This is a margin question, not a philosophy question. It can be calculated.
Defining the comparison properly
FBA (Fulfillment by Amazon) means Amazon stores, picks, packs, ships and handles customer service and returns. You pay a per-unit fulfillment fee based on size tier and weight, plus monthly storage on cubic feet, plus surcharges for aged inventory.
FBM (Fulfillment by Merchant) means you or your 3PL ship to the customer. You pay no FBA fulfillment fee and no FBA storage fee. You pay actual shipping, packaging, labour, warehousing and the cost of handling returns yourself.
The referral fee is the same either way. That is the point most comparisons miss: the choice is not about Amazon’s cut, it is about who performs the logistics and at what cost.
Two things also change that are not fees:
- Prime eligibility. FBM is only Prime-eligible through Seller Fulfilled Prime, which carries strict performance requirements. Without it, conversion rate is lower.
- Buy Box competitiveness. FBA offers are advantaged in Buy Box logic. An FBM offer generally needs to be priced lower to win the same share.
Both of those are real costs and belong in the model.
The comparison framework
Compare on contribution per unit, then sanity-check on contribution per dollar of working capital.
FBA contribution = Price − Referral − FBA fulfillment fee − Storage allocation
− Returns cost − Landed cost
FBM contribution = Price − Referral − Outbound shipping − Packaging
− Pick/pack labour − Warehousing allocation
− Returns cost − Landed cost
Then adjust for the conversion gap:
Effective FBM contribution = FBM contribution × (FBM conversion / FBA conversion)
That ratio is the number sellers refuse to estimate, and it is where the decision usually turns. If FBM converts at 70% of FBA for the same SKU at the same price, FBM needs a 43% higher per-unit contribution just to break even on volume.
Worked example: where FBM wins
Illustrative figures. Substitute your own.
SKU A — large, heavy, slow-moving. 22 lb, oversize tier, sells 15 units/month, price $89.
| Line | FBA | FBM |
|---|---|---|
| Price | $89.00 | $84.00 |
| Referral (15%) | −$13.35 | −$12.60 |
| Fulfillment | −$16.90 | — |
| Outbound shipping | — | −$11.40 |
| Packaging + labour | — | −$2.10 |
| Storage / warehousing | −$3.80 | −$0.60 |
| Returns allowance (4%) | −$1.30 | −$1.60 |
| Landed cost | −$38.00 | −$38.00 |
| Contribution | $15.65 | $17.70 |
Adjust FBM for a 0.80 conversion ratio: $17.70 × 0.80 = $14.16 effective.
FBA still wins here — narrowly. But note what happens if this SKU sits long enough to attract aged-inventory surcharges. At 15 units/month with a 90-unit shipment, that is six months of cover. Add a modest aged surcharge and FBA storage climbs while FBM warehousing barely moves. The ranking flips.
SKU B — same SKU, but you already run a warehouse for another channel. Marginal warehousing and labour cost, not fully allocated: storage $0.20, labour $0.90.
FBM contribution: $89.00 − $13.35 − $11.40 − $1.10 − $0.20 − $0.90 − $1.60 − $38.00
= $22.45 (priced at parity, accepting lower conversion)
Effective: $22.45 × 0.80 = $17.96 vs FBA $15.65
FBM wins by 15%, entirely because the fixed logistics cost was already being paid for another reason.
The SKU profile where FBM tends to win
- Oversize and heavy. FBA fulfillment fees scale aggressively with size tier and weight. Your own negotiated freight often scales better.
- Slow-moving. Storage is charged on time and space. Low turnover is expensive in FBA and cheap in a warehouse you already have.
- Low unit value with awkward dimensions. Where the fulfillment fee is a large percentage of price.
- Highly seasonal. FBA capacity constraints and peak surcharges hit hardest exactly when you need the inventory in place.
- Fragile or requiring special handling. Where FBA damage rates raise your returns cost above what careful in-house packing would.
- Bundles and multipacks assembled to order. Prepping variations for FBA multiplies SKU count and inventory risk.
- When you already operate fulfillment. The marginal cost argument above is the strongest case there is.
The SKU profile where FBA is not close
- Small, light, fast-moving, price-competitive items where the fulfillment fee is a few dollars and conversion is everything.
- Anything where you are competing for the Buy Box against multiple FBA sellers at a similar price.
- Categories with high return rates where Amazon absorbing the customer service load has real operational value.
Real-world implications beyond the fee math
Working capital. FBA requires inventory to sit in Amazon’s network before it can sell, and capacity limits may cap how much you can send. FBM lets you hold stock centrally and allocate on demand. For a cash-constrained seller, that flexibility can be worth more than a per-unit fee difference.
Performance risk. FBM puts on-time delivery and valid tracking rate on your account. Miss them and you have an account health problem, not just a margin problem. This is the asymmetry that makes FBM genuinely riskier: the downside isn’t lower profit, it’s suspension exposure.
Hybrid is normal. Nothing requires one method per account. Running FBA on fast movers and FBM on oversize slow movers is the answer more often than either extreme, and it is what the math usually points to.
Best practices
- Run the comparison per SKU, never per account. The answer varies by size tier and velocity, so an account-level policy is guaranteed to be wrong for part of the catalogue.
- Estimate the conversion gap explicitly. Even a rough number beats leaving it out. Test it by running one SKU both ways for a month.
- Use marginal cost only if it’s genuinely marginal. If moving a SKU to FBM means hiring someone, that is not marginal.
- Re-run the comparison when fee schedules change and after any dimension remeasurement.
- Include aged-inventory surcharges in the FBA side for anything with more than 90 days of cover.
- Track net margin per SKU per fulfillment method so the decision is reviewed on evidence, not on the assumption you made at launch.
Tools like sellerboard track net profit at SKU level with fulfillment costs separated out, which is what makes a same-SKU comparison across methods possible after the fact rather than only in a spreadsheet beforehand.
Common mistakes
- Comparing FBA fees against outbound shipping only. Packaging, labour, warehousing and returns handling are real and routinely omitted, which flatters FBM.
- Ignoring the conversion gap. The most common error in the other direction. An FBM offer without Prime does not sell at the same rate.
- Allocating full overhead to a marginal FBM decision. If the warehouse exists regardless, loading it with full absorbed cost buries a genuinely profitable option.
- Forgetting FBM return shipping. Customers return to you, and you pay for it.
- Underestimating time cost. Fulfillment work you do yourself is not free just because it isn’t invoiced.
- Treating the decision as permanent. Velocity changes, fees change, seasonality changes. This is a quarterly review, not a launch decision.
FAQ
Is the referral fee different for FBA and FBM? No. Referral fees are category-based and identical. Only the fulfillment side differs.
Can FBM offers be Prime-eligible? Only through Seller Fulfilled Prime, which requires meeting strict delivery speed and performance standards. Without it, FBM offers are not Prime-eligible.
How much lower does FBM convert? It varies widely by category, price point and competition. Rather than adopt a benchmark, test the same SKU both ways for 30 days and measure it in your own account.
Can I use both methods for the same SKU? Yes, and many sellers do, holding FBA stock for Prime demand and FBM as an overflow or backup during stockouts.
Does FBM hurt account health? It moves delivery performance onto you. Handled well, it is neutral. Handled badly, it creates account-level risk that FBA would have absorbed.
Conclusion
FBA versus FBM is not a strategic identity, it is a per-SKU cost comparison with two adjustments most sellers skip: the conversion gap and the marginal-versus-absorbed cost question.
Run it on your five largest and five slowest SKUs. On the fast small items you will almost certainly confirm what you are already doing. On the oversize slow movers you may find a margin you have been paying Amazon to take care of for you — and deciding whether that convenience is worth its price is a reasonable thing to do deliberately rather than by default.
