A distributor that publishes an ROI on every line is doing you a favour and setting a trap at the same time. The favour is obvious: you can rank a catalog by expected return instead of running each product through a tool. The trap is that a single percentage hides five inputs, each of which can be wrong for your account in particular. This guide explains what a per-product ROI is made of, how to read it, and where it stops being useful.

The five inputs

  1. Your cost. The trade price per unit, plus anything it takes to get the unit sellable: inbound freight, prep, labelling. A catalog ROI usually starts from the trade price alone.
  2. The selling price. Usually the current Buy Box price on Amazon.com, sometimes a recent average. Which one matters: a Buy Box briefly inflated by a stockout makes every line look better than it is.
  3. Amazon's fees. The referral fee (a percentage of the selling price that varies by category), the FBA fulfilment fee (set by size and weight tier), and for media categories such as books, a fixed closing fee. Storage, inbound placement and returns processing come on top.
  4. Demand. Sales rank and an estimated monthly sales figure. Neither is a promise; both tell you whether the price is one the market actually pays.
  5. Competition. The number of sellers on the listing and whether Amazon itself is one of them. This rarely appears in the ROI number but decides how much of the demand you will actually get.

ROI is then simply (selling price − fees − cost) ÷ cost. Unit profit is the same calculation without the division. Both are only as good as the inputs.

Where the inputs come from

The trustworthy sources are well known. Keepa records Buy Box price and sales rank history for Amazon listings and derives a sales estimate from the rank curve. Helium 10's Xray tool produces its own sales estimates. Amazon publishes its fee schedule and, through the Selling Partner API, will return a fee estimate for a specific listing at a specific price. A catalog that names its sources is one you can check; a catalog that shows a bare “35% ROI” with no provenance is asking you to trust it.

Family-level and child-level sales

Many Amazon listings are variation families: one parent with children for size, colour or pack count. Sales rank is shared across the family, so a rank-based estimate describes the whole family, not the child you are buying. A tool that measures the child directly can return a far smaller figure for the same ASIN, and both can be right. When a distributor shows a family figure against a child product, the demand for your exact variation may be a fraction of it. Ask which level the estimate describes; if the catalog does not say, assume family.

The reverse is also useful. A child that sells little may sit in a family whose best-selling sibling moves hundreds of units a month. If the distributor can supply that sibling, the “dead” line is an opportunity rather than a warning.

What Athena shows you

Lines in the US catalog show the trade price in USD beside the Amazon.com figures for the same product wherever that data is available: the current Buy Box price, the referral and FBA fees, the sales rank, an estimated monthly sales figure, and the estimated unit profit and 90-day ROI that follow. The sources are Keepa for Buy Box and rank history, Helium 10 exact-child estimates where an exact match exists, and Amazon's own fee estimates through the Selling Partner API, refreshed on a daily job.

Two rules keep the figure honest. First, where a sales estimate can only be measured at family level, the catalog labels it as the family figure rather than presenting it as the child's. Second, when the best-selling sibling in a family is a different product that measures at least 50 sales a month, the line carries an “Opportunity” marker so you can ask for the variation that sells. And where the source data is missing, the line shows nothing: Athena never substitutes an invented number.

Why the ROI on a website is not the ROI you earn

  • Your fees are not the catalog's fees. Referral fee tiers, FBA storage, inbound placement, low-inventory fees and returns vary by account and by month. Add your own.
  • Prep and inbound are real. Poly bags, labels, freight to the fulfilment centre. A 40% catalog ROI on a $6 item can be a 20% ROI once the unit has been bagged and shipped.
  • The Buy Box moves. Model the price you expect to sell at in the month your stock arrives, not the price on the day you looked.
  • Demand is shared. Monthly sales are for the listing; your share depends on how many sellers are on it and whether Amazon is one of them.
  • Estimates are estimates. Rank-based sales estimates are curves fitted to observed data. They are useful for ranking and useless for guarantees.

A short buying discipline

  1. Filter the catalog by ROI and monthly sales to make a shortlist; do not buy from the filter.
  2. Open each line and read the Buy Box history, the rank and the seller count yourself.
  3. Recalculate the ROI with your own fee tier, prep cost and inbound freight.
  4. Check the child-versus-family question for anything with variations.
  5. Buy repeatable lines in quantities you can sell through, then replenish.

The live demo shows the real numbers on real lines exactly as an approved account sees them, which is the fastest way to see how the inputs above are presented before you apply.