Your ACOS won't come down, and your first instinct is to cut bids. Hold on. I've looked at enough seller ad accounts over the years to see the pattern: when ACOS refuses to move, the problem is usually somewhere other than the bid. Here are three checkpoints, in order. Most sellers find the root cause by the second one.
First, get the math straight before you optimize anything. ACOS is ad spend divided by ad-attributed sales. Your break-even ACOS equals your gross margin — above that line, every ad-driven order loses money. Take a $40 listing as an example: Amazon's referral fee is 15% in most categories, that's $6; FBA fulfillment fees follow whatever Amazon currently publishes, call it $8 for this exercise; assume a $10 product cost and $2 of inbound freight per unit. Gross profit is $40 − $6 − $8 − $10 − $2 = $14, a 35% margin, so break-even ACOS is 35%. A keyword sitting at 60% ACOS means you're paying out of pocket on every order it generates. That math is the foundation for all three checkpoints below.
Checkpoint one: are your search terms burning money for nothing?
Pull the search term report in your ad console, last 30 days, sorted by ACOS descending. Look for two kinds of terms: ones spending steadily with ACOS above your break-even line, and ones that spent money and produced zero orders.
Don't rush to kill the first kind. Check the conversion rate first. If it's at or above the category average, the keyword is relevant — it's just overpriced. Lower the bid one notch and watch it for two weeks. Pausing it outright throws away the conversion history it built, which is a worse deal. If the conversion rate is also poor, the term was never a match for your listing; no bid adjustment will fix that. Negate it and move on.
The second kind — money spent, zero orders — gets exact-match negated, no mercy. Auto campaigns are full of these "surprise" terms. Last year I reviewed an account for a pet supplies seller in Ontario whose auto campaign was carrying more than twenty terms that had each burned over a hundred dollars with nothing to show for it. Negating them cut the account's overall ACOS by nearly ten points in two weeks. Money isn't saved; it's stopped from bleeding.
Negation has its own craft, in three layers. Layer one is campaign-level negatives for obviously irrelevant terms — block them once, across the whole campaign. Layer two is ad-group-level exact negatives for terms that already spent money and proved they don't convert. Layer three is the one most sellers miss: pull brand terms out and look at them separately. Traffic searching your brand name converts naturally high with naturally low ACOS. Don't hold broad generic keywords to your brand term's ACOS, or you'll misjudge every generic term as "burning money" and stop bidding on terms you should be winning. Review the negative list once a week, thirty minutes — the highest-ROI half hour in ad optimization.
Checkpoint two: did a stockout break your ad momentum?
This is where the most sellers get hurt, and it's the tightest link between advertising and warehouse operations. When FBA runs out of stock, the listing's rank slides. When inventory returns and ads ramp back up, ACOS doubling is normal. Amazon's logic is unsentimental: your conversion rate was zero during the stockout, so why would the system keep giving you cheap traffic?
Replenishment math belongs on the same spreadsheet as the ad budget. My formula is unglamorous but it works: daily unit sales × (inbound transit time + FBA receiving time) × a 1.5 safety factor. Don't guess transit time — ask your forwarder for the actual average of the last three months. FBA receiving time follows Amazon's latest published policy, and you add buffer in peak season. The 1.5 factor is my own rule of thumb, reserved for demand spikes and logistics delays.
Run the numbers once and the scale of the problem becomes obvious. Say a listing sells 50 units a day, ocean transit runs 35 days (get the real number from your forwarder; this is an assumption), and FBA receiving takes 10 days (per Amazon's latest published policy). Safe stock is 50 × 45 × 1.5 = 3,375 units. Plenty of sellers plan only "in transit plus on hand" and quietly drop the 1.5 — then one demand spike and they're out. Also watch Amazon's FBA storage capacity and restock limits (per Amazon's latest published policy) and request more capacity before peak season, not after you discover shipments can't get in.
And when you come back from a stockout, don't restore the original budget on day one. Run at half budget for a week, let the conversion rate recover, then scale back up gradually. Ranking weight is earned back, not bought back — slamming the budget to full on day one just buys you a higher ACOS.
The ugliest way to die is ramping up ads before a big promo while the inventory is still on the water. Ad spend burns on schedule, shoppers click through to an "out of stock" page and leave, conversion rate collapses, ACOS explodes, and the ranking you paid for goes with it. My iron rule: staying in stock is the precondition for advertising. Bids can be tuned slowly. One stockout wipes out at least half the value of three months of ad spend. An operator should check the replenishment sheet more often than the ad report.
Checkpoint three: are you reading the ad P&L or the business P&L?
ACOS only counts ad-attributed orders, but your business is more than ad orders. TACOS — ad spend divided by total sales — is the number that shows what advertising really costs your whole business. ACOS alone will lie to you: 25% ACOS looks great, but if organic orders are being cannibalized by ads, total profit may be shrinking.
Here's how to tell real improvement from fake: TACOS trending down while total sales trend up — that's advertising making money. ACOS down but TACOS flat or rising — you've just converted organic traffic into paid traffic and called it a win. An example: a listing does $10,000 a month, ads drive $4,000 of that on $1,000 of spend — ACOS 25%, TACOS 10%. Next month ad spend is $1,200, ad sales $4,800, ACOS still 25%, but total sales reach $14,000 and TACOS drops to about 8.6%. ACOS unchanged, TACOS down, total sales up — that is real optimization. Numbers lie; read them side by side and they can't fool you.
One more case deserves its own note: during a new product launch, ACOS above break-even is normal. A launch needs rank and reviews, and plenty of veteran sellers will knowingly lose money on ads for three months to build a listing that can fight. The condition is that you know exactly how much you're losing and for how long, with a monthly review — not burning blindly. The end goal of advertising isn't a pretty ACOS. It's that every dollar spent buys incremental growth.
Do one thing this week; it takes thirty minutes. Download the last 30 days of search term reports and exact-negate every term with ACOS above your break-even line and zero orders. Then don't stare at it daily. Pull the report again at the same time next week and compare overall ACOS and TACOS. Ad optimization is slow work. One check a week is plenty.
One question to leave you with: is your reorder alert set from sales math, or from gut feeling? If it's the latter, your second task this week is plugging the replenishment formula above into your spreadsheet.
