How to evaluate an Amazon wholesale opportunity
An Amazon wholesale brand is worth pursuing when three of its ASINs clear 10% net ROI on $1,000 a month of revenue and face three or fewer other sellers.
Last reviewed 5 min readSupplyForge LLC
An Amazon wholesale opportunity is worth pursuing when three of the brand's ASINs clear 10% net ROI after Amazon fees on $1,000 or more of monthly revenue each. Seller count then decides how much of that revenue is actually available to you.
The test runs in two stages, and the order matters. Hard filters go first because they are cheap, mechanical, and end the conversation. Scoring signals run only on the survivors, because judgment is the expensive input.
Stage one: the hard filters that end the conversation
Hard filters are pass-or-fail and applied before any human looks at a candidate. Forge Command runs this set before surfacing anything in its sourcing module, and every threshold below is a literal number in that filter.
| Filter | Threshold | Why it exists |
|---|---|---|
| Net ROI after Amazon fees | At least 10% | One fee change puts a thinner SKU underwater |
| ASIN monthly revenue | At least $1,000 | Research time is the scarce input |
| Profitable ASINs on the brand | At least 3 at 10%+ ROI | A one-SKU brand does not repay opening an account |
| Multipack listings | Excluded | Price is per unit, the listing per pack, so ROI is wrong |
| Generic or empty brand | Excluded | No brand means no supplier to open an account with |
| Supplier forbids third-party Amazon sales | Excluded | The account cannot legitimately supply Amazon |
The three-ASIN rule saves the most time: opening a wholesale account costs the same effort whether the line yields one profitable SKU or twenty. Multipacks are excluded for the reason set out in why multipack listings break wholesale margin maths.
Stage two: margin bands and velocity tiers
The 10% floor is a survival line, not a target. Above it, net ROI and monthly revenue sort into bands with different verdicts, and the top margin band is a warning rather than a reward.
Net ROI bands: 10–14% barely viable, and only when demand is strong and competition low; 15–24% comfortable, the floor for a routine yes; 25–34% strong; 35–49% very strong. Above 50%, investigate first — an outlier margin is more often a data error than a find: the wrong ASIN matched to the SKU, a missing prep cost, or a MAP breach.
Revenue tiers per ASIN: $10,000 a month or more is ideal, $5,000 to $10,000 strong, $1,000 to $5,000 viable only when margin and competition are both favorable. Unit counts cross-check the tier — above 300 a month is very strong, 100 to 299 strong, 30 to 99 viable, below 30 marginal. A 90-day trend above 1.1 is rising, below 0.9 declining. The cross-check matters: $9,000 at 30 units is a high-ticket item one new seller halves, while $9,000 at 400 units absorbs an entrant.
Jungle Scout's self-reported survey data puts 58% of wholesale sellers under a 20% profit margin and 18% between 26% and 50%; those pages are dated 2024, so treat them as directional. The full calculation is in how to calculate net ROI on a wholesale product.
Competition: three seller counts kept side by side
Seller count is three different numbers measuring three different things, and merging them destroys the information. Forge Command stores all three separately and never reconciles them.
- SP-API live offer count — new-condition offers on the listing right now. Total sellers, not FBA-only.
- Helium 10's seller count — a snapshot as of the last CSV upload, not a live reading.
- Keepa's buy-box seller count — distinct buy-box winners over 30 days. Structurally lower: a seller who is listed but never wins does not appear.
The bands: one seller is best, two still good, three workable, four to five needs strong margin and demand, six or more usually a skip. See how many sellers on an ASIN is too many.
Two named patterns override the raw count. An FBM buy-box winner with three or fewer sellers is an entry signal, per the FBM buy box pattern, while a lone FBA seller who is the brand itself is a no-go, per when the brand is the only FBA seller.
Worked illustration: scoring one brand end to end
The numbers below are a worked calculation for a hypothetical brand, not a customer result. Eleven ASINs in the catalog; four clear both floors, so the brand passes the three-ASIN threshold. The flagship SKU:
- Wholesale cost $27.50, plus $1.10 inbound and prep, giving a landed cost of $28.60.
- Amazon price $49.99. Referral fee at 15% is $7.50; FBA fulfillment $6.35. Total Amazon fees $13.85.
- Net proceeds $36.14. Profit per unit $7.54. Net ROI 26.4%, net margin 15.1%.
- Listing revenue $9,498 a month at 190 units. SP-API shows three sellers; Keepa shows two distinct buy-box winners over 30 days.
Now the step most people skip. With three sellers rotating, model your share at roughly a third: 63 units, about $475 a month from this SKU, and roughly $1,900 across the four qualifying ASINs. That $1,900 is the number that has to justify the account — not the $9,498 of listing revenue.
The tier verdict and what to do with each
A candidate that survives the filters gets one of three tiers, and each tier maps to a timeframe rather than to a score.
| Tier | What it means | Action |
|---|---|---|
| Top | Margin 25%+, strong velocity, three or fewer sellers, no restriction flag | Pursue this week |
| Strong | One signal merely acceptable — 15–24% margin, four to five sellers, or exactly three qualifying ASINs | Pursue this month |
| Marginal | Several middling signals, or one real concern such as stale data or a declining trend | Research first |
Anything weaker than marginal should not be surfaced at all. The worked illustration lands at strong rather than top: the margin band qualifies, but three sellers and a 60-day-old Helium 10 snapshot each cost it a notch.
Where the inputs come from and when they go stale
The scoring framework consumes data it does not generate, and every input has a shelf life. Forge Command reads velocity from Helium 10 CSV uploads, price and buy-box history from Keepa, and live pricing, fees, and offer counts from SP-API in batches of 20 identifiers per request.
Three honest caveats. Forge Command does not replace Helium 10 or Jungle Scout for research — it consumes their output. Keepa lookups are opt-in and spend the tenant's own paid tokens, and Keepa plan pricing could not be verified on a primary source as of 2026-08-03 because the page sits behind Cloudflare. A Helium 10 upload older than about 90 days should be re-pulled before you act on its seller count.
SP-API itself is free as of 2026-08-03, and Amazon commits to at least 90 days' notice before a fee increase — the planning window your ROI floor needs headroom for.
Frequently asked questions
Why is the ROI floor 10% and not 15%
Ten percent is the exclusion line, not the target. It removes candidates that cannot survive a fee change; the band that reads as a routine yes in this framework is 15% to 24% net ROI after Amazon fees.
Should I divide expected sales by the number of sellers
Yes, as a default. Splitting listing volume by seller count is a rough model rather than a rule, but it beats assuming you take the whole listing, and it is the difference between a $9,498 opportunity and a $475 one.