Whether a keyword pays off is decided by the distance between its ACOS and its own break-even ACOS — and that second number is different for every keyword in your account. Break-even ACOS is not a property of your business or even of a product. It is a property of a row: it depends on what that keyword actually sold, at what price, with which promotions attached, how many of those orders came back as refunds, and what fee profile the units carried. Two keywords sitting at exactly 21% ACOS can land on opposite sides of zero. The PPC Dashboard in sellerboard calculates a separate break-even ACOS for every portfolio, campaign, ad group, keyword and search term, which is why “is this keyword making money?” is usually a column you read rather than a calculation you run.
Why does a keyword with steady sales still lose money?
ACOS treats revenue as the finish line. Profit keeps going past it. Between a PPC-attributed sale and the money you keep sit six more deductions: promotions applied to the order, refund costs, referral fee, FBA fee, other Amazon fees, cost of goods, and any indirect expenses allocated to the period. Ad spend is one line in that stack, not the stack.
This is exactly what the summary panel to the right of the chart shows. It is a profit-and-loss statement for PPC-attributed sales only, in this order: PPC sales, Promo, Ad spend, Refunds, Amazon fees (expandable into FBA fee, referral fee and other Amazon fees), COGS, Expenses, and Profit (estimated). A keyword can clear the ad-spend line comfortably and still fail somewhere in the five lines below it.
What does the Break even ACOS column actually measure?
sellerboard computes it per row:
Break even ACOS = (Profit estimated + Ad spend) ÷ PPC sales × 100%
Read it in two steps. Adding ad spend back to profit reconstructs what that keyword earned before advertising — its contribution after product costs, fees, refunds and promotions. Expressing that as a share of PPC sales tells you what proportion of revenue could have gone to ads with exactly nothing left over. Compare it with the ACOS column and the verdict is immediate: below break-even, the keyword banks the difference; above it, the keyword bills you the difference.
The gap between the two percentages is not just a signal, it is the amount. Multiply the gap in percentage points by the row’s PPC sales and you get the money the keyword made or cost you over the period.
What does an unprofitable keyword look like in the table?
Here is a keyword with 30 days of clean, unremarkable performance: 332 clicks, 41 orders, $1,480 in PPC sales, $312 in ad spend. Its ACOS is 21.1%, which sits well under the 24% break-even the seller worked out for the product back when they set the campaign up.
| Line | Amount |
| PPC sales | $1,480.00 |
| Promo | −$96.00 |
| Refunds | −$104.00 |
| Amazon fees (FBA + referral + other) | −$488.00 |
| COGS | −$460.00 |
| Expenses | −$45.00 |
| Contribution before advertising | $287.00 |
| Ad spend | −$312.00 |
| Profit (estimated) | −$25.00 |
Now run the two percentages:
| Metric | Arithmetic | Result |
| ACOS | $312 ÷ $1,480 × 100 | 21.1% |
| Break even ACOS | (−$25 + $312) ÷ $1,480 × 100 | 19.4% |
| Gap | 21.1% − 19.4% | 1.7 points |
| Cost of the gap | 1.7% × $1,480 | $25 |
The keyword sells. It converts. It looks like a keeper on every Amazon Ads report you could pull. It is also quietly $300 a year of negative margin, and the reason is 1.7 percentage points that only exist once promotions, refunds and real COGS are in the calculation.
The dashboard also gives you the corresponding bid: Break Even Bid = Break even ACOS ÷ ACOS × CPC. At a $0.94 CPC, that is 19.4 ÷ 21.1 × $0.94 = $0.86. Treat it as a boundary marker rather than a target — sellerboard flags it as theoretical, and bidding at exactly break-even usually costs you a chunk of impressions for zero profit.
Why do two keywords at the same ACOS end up on opposite sides of zero?
Because ACOS knows the revenue and the spend, and nothing else. Here are two keywords from the same ad group, both at 21.1% ACOS over the same 30 days:
| Metric | Keyword A | Keyword B |
| PPC sales | $1,480 | $1,510 |
| Ad spend | $312 | $319 |
| ACOS | 21.1% | 21.1% |
| Break even ACOS | 19.4% | 28.6% |
| Profit (estimated) | −$25 | +$113 |
| Margin | −1.7% | 7.5% |
Four things routinely drive that spread, and the dashboard exposes all of them:
- The keyword sells a different product than the one you advertised. The Same SKU / All SKUs column compares sales of the advertised product against total sales that followed the click. A keyword whose revenue leaks into a cheaper variant is being judged on the advertised product’s economics while it is really funding another one.
- Promotions. Coupons and deals land on the Promo line, not on the ACOS. A 10% coupon running on one SKU moves its break-even ACOS by roughly the same margin it removes.
- Refunds. A keyword that attracts marginal buyers — wrong size, wrong expectation, comparison shoppers — carries a higher refund rate, and the refund takes the fees and the unit with it.
- Fee profile. Two SKUs at the same price can differ by a dollar or more per unit in FBA fees alone, and low-price-point units are the ones where a dollar decides the outcome.
How do I pull the list of keywords that sell but don’t pay off?
The PPC Dashboard can produce exactly that list in about a minute:
- Set the date range to Last 30 days, by day, or Last 3 months, by week if your keywords are low-volume and a month of data is too thin to judge.
- Open the Keywords tab — or stay on Ad groups and expand rows with the › arrow to see keywords in context.
- Use the Columns panel (grid icon) to switch on Profit, Margin, Break even ACOS, Break Even Bid, Cost per order and Same SKU / All SKUs. Keep the ACOS heatmap on: green to red tells you where to look before you read a single number.
- Open More filters and set two numeric ranges: Sales minimum at whatever counts as meaningful volume for you, and Profit maximum at 0. What comes back is the definition of the problem — keywords with real sales and no profit.
- Sort by Profit ascending to rank them by how much they are actually costing you, not by how bad their ACOS looks.
- Export to CSV with the download icon if you want to work the list outside the app; the export respects the active tab and every filter.
The All keywords filter also has a Show bad performing keywords only option for a faster, less surgical pass over the same ground.
Lower the bid or pause the keyword?
The keyword’s own numbers usually answer this. Compare Current bid with Bid recommendation: the recommendation is derived from your break-even analysis and the profit goal configured for that campaign, so the distance between the two tells you whether this is a small correction or a structural problem.
- Small gap, healthy conversion: bring the bid to the recommendation. On the Campaigns tab, Accept recommendations applies them in bulk.
- Break-even bid far below what wins impressions: the keyword is priced out of profitability at the current unit economics. Fixing it means moving COGS, price, promo depth or fee profile, not the bid.
- Same SKU / All SKUs badly skewed: the traffic is fine, the targeting is wrong. Point it at the product it is actually selling and it can flip to profitable at the same CPC.
- Low volume, negative every period: pause it and put the budget behind a keyword whose break-even ACOS has headroom.
One more place to look before you touch the bid: the Search terms tab. A keyword can be profitable in aggregate while two or three of the queries matching it drain the rest. The tab is read-only in sellerboard because search terms cannot be edited — but once you know which ones they are, you negate them in Amazon and the keyword’s economics change without touching a bid.
Common mistakes
- Using one account-wide break-even ACOS as the benchmark for every keyword, when the dashboard calculates one per row.
- Judging a keyword on ACOS alone and never adding the Profit and Break even ACOS columns to the table.
- Recalculating break-even once a year. Supplier price changes, fee updates, a new coupon or a rising refund rate all move it, and the campaign shows no sign of it.
- Pausing everything above the target ACOS. Some of those keywords have a break-even ACOS of 35% and are earning well.
- Bidding at the Break Even Bid on purpose. It is the point where profit reaches zero, and it usually costs impressions on the way there.
- Ignoring the Same SKU / All SKUs column, then wondering why a “bad” keyword is somehow correlated with sales of another product.
- Judging a low-volume keyword on 30 days. Widen the range before you act.
FAQ
Is the profit shown per keyword exact?
It is labelled “estimated” for a reason. PPC sales, orders, clicks and ad spend come from Amazon; refunds, promotions and Amazon fees are estimated per PPC order from the previous month’s averages for those products, and COGS comes from your product data. The estimate is close enough to rank keywords against each other and to identify the ones that are underwater — that is what it is built for.
Why is my keyword’s break-even ACOS different from my product’s break-even ACOS?
Because a keyword rarely sells one product in one clean configuration. The orders behind it can include other SKUs bought after the click, discounted units, multi-unit orders and refunded orders. The row-level break-even ACOS reflects what actually happened; the product-level figure reflects a single idealised unit.
My ACOS improved and my profit fell. How?
Look at what the ads sold, not just what they cost. A shift toward a lower-margin SKU, a coupon that went live mid-period, or a jump in refunds can push break-even ACOS down faster than you pushed ACOS down.
Can I see this at ad group or campaign level too?
Yes. Break even ACOS, Profit and Margin are calculated on every tab — Smart portfolios, Portfolios, Campaigns, Ad groups, Keywords and Search terms. Campaign level is the fastest way to find where to dig; keyword level is where the fix usually is.
Does this work outside Sponsored Products?
The campaign type filter covers Sponsored Products, Sponsored Brands, Sponsored Brands Video, Sponsored Display and RAS, and the profitability columns apply across them.
A keyword’s job is not to produce sales, and not to produce a low ACOS. Its job is to buy revenue for less than that revenue is worth after everything Amazon, your supplier and your buyers take out of it — and that threshold is different for every keyword you run. Once break-even ACOS sits next to ACOS in the same table, the keywords that sell but don’t pay off stop hiding behind their order count.
