The standard mistake on a stuck TACoS account is to cut spend. The brand sees ad costs climbing as a share of revenue, gets nervous, and pulls the lever that feels safest. Spend comes down. Revenue comes down faster. TACoS gets worse. The account is now both smaller and sicker.
This is the workflow I run instead. It is deliberately simple. The point is not to look smart in a deck. The point is to find the actual problem on the account in one or two sittings and fix only that.
Step 1: pull the bulk report on the right window
Download a Sponsored Products bulk report. Window length depends on the account.
- Small or quiet account: 30 days of data.
- Medium account: 14 to 21 days.
- Big or fast-moving account: 7 days.
Anything beyond 30 days is too much. The market moves, your product moves, your inventory moves. Decisions made on 90-day-old data are decisions made on a different account.
Step 2: bucket every keyword by ACoS range
Tag each keyword into an ACoS bucket. Use plain ranges, do not over-engineer:
0 to 10%, 10 to 20%, 20 to 30%, 30 to 40%, 40 to 50%, 50 to 60%, 60 to 70%, 70 to 80%, 80 to 90%, 90 to 100%, 100 to 200%, and 200%+.
That is the X-axis. Now you can see your account as a distribution, not a single TACoS number.
Step 3: lay spend% and sales% over the buckets
For each ACoS bucket, calculate two things: what percentage of total ad spend sits in that bucket, and what percentage of total ad-attributed sales sits in that bucket.
This is the only number that matters for the next step. Two columns. One sheet. No dashboards needed.
Step 4: calculate Max Paid ACoS from CM2
Run it on the account’s blended CM2 target. Lock the number. Example: it lands at 70%. Anything above that line is losing money on paid traffic.
Now read the distribution from Step 3 against this line. Two cases:
- Majority of ad-attributed sales come from buckets ABOVE Max Paid ACoS. Account is broken. The fix is not bid-down. The fix is in targeting, products, or campaign architecture. Bulk operations will not save you. Investigate one issue at a time and fix it manually. Wrong audience? Wrong intent on the keywords? Listings not converting? Pricing out of category? Take them one by one.
- Majority of ad-attributed sales come from buckets BELOW Max Paid ACoS. Account is healthy. You can run the bid-down ladder in Step 5.
This is the diagnostic. Two cases, one decision.
Step 5: the bid-down ladder (healthy account)
If the distribution is healthy, walk down the high-ACoS buckets and trim:
Size each cut to how far the keyword sits over Max Paid ACoS. Apply against a 7 to 14 day ACoS window so you are not reacting to one bad day.
For raises on the other side: any keyword sitting comfortably under Max Paid ACoS with strong impression share signal, raise bids by 5 to 10% sized to how far under target the ACoS sits, judged on a 30 to 60 day window. Slow and patient on the way up. The raises compound; the cuts protect.
Step 6: discovery keywords, auto campaigns, and the loop
Discovery phase keywords (no sales yet). Bucket by clicks: under 5, under 10, under 20. How much you invest in discovery depends on inventory health and your CM2 goal.
- If margin and inventory are tight: cut bids by 5% on anything above 5 clicks zero-sales. You cannot afford the bleed.
- If margin and inventory are healthy: run the CVR-anchored negation rule. Math first: at ~10% CVR, expect ~10 clicks per sale; halve that to 5% for benefit of doubt → ~20 clicks per expected order.
Auto campaigns are not part of this workflow. Do not bulk-edit Auto. If an Auto campaign is misbehaving, investigate manually. Bid multiplier wrong? Seasonal search-term spike? Targeting drift? Find the actual reason before you touch the bid, because Auto is A9-aligned and easy to break.
Run negatives regularly. Do not skip this step. Pull the search terms report on the same cadence as the bulk report. Anything passing the click thresholds with no conversion goes on the negative list. Anything category-wrong gets phrase-negated.
The loop, in order. Bids down → check search terms → run negatives → check placement reports → fix where the actual issue is. Not all four at once. One at a time.
Cadence by account size.
- Big account, fast spend: weekly.
- Medium account: every two weeks.
- Small or seasonal account: monthly.
That is the workflow. Find the issue, solve it, do not over-engineer the rest. If you want a written diagnosis on your specific account before running it yourself, the Amazon Account Audit is the right scope.
Bottom line
Cutting TACoS without killing revenue is not a budget exercise. It is a diagnostic exercise. Pull the bulk, bucket the ACoS, lay spend% and sales% against Max Paid ACoS, decide whether the account is broken or healthy, then run the right move – manual investigation or the bid-down ladder. Run the loop on a cadence. Stop cutting top-line spend by reflex; you cannot grow an account out of a problem you have not diagnosed.
If you want me to walk through your account against this diagnostic, message me on WhatsApp.