Sample Ads Diagnostic — read what $497 buys before you spend it.
This is a condensed, anonymized sample of the Amazon Ads Diagnostic. The client version runs 15–25 pages and is built from your bulk reports; this one is a composite teaching case — a realistic account assembled from patterns I've found across real accounts over 4 years, with representative numbers. Every finding below is a real, recurring pattern. If several feel familiar, that's the point.
1 · Account scorecard
The one-sentence verdict: this account pays $2,731 every month — 28% of its ad budget — for clicks that have already proven they don't convert, while its three most profitable campaigns run out of budget by early afternoon. The account isn't under-optimized; it's mis-allocated.
Client version adds: per-SKU scorecards, placement breakdown, 12-month trend, and category benchmarks for your ACOS band.
2 · The findings
1The auto campaigns became the account
Auto campaigns were meant to be discovery — two years later they're 38% of all spend and nobody has harvested them. The 14 search terms below converted repeatedly inside auto but were never promoted to exact-match campaigns, so the account keeps paying auto-tier CPCs for demand it already discovered. Sample of the harvest list:
| Search term (in auto) | 30-day spend | Orders | ACOS | Should be |
|---|---|---|---|---|
| [hero product] with lid | $412 | 31 | 14.2% | Own exact campaign |
| [hero product] dishwasher safe | $286 | 19 | 17.8% | Own exact campaign |
| [SKU-3 product] small kitchen | $204 | 12 | 19.5% | Own exact campaign |
Client version: the full harvest list (typically 10–25 terms), each with its destination campaign, suggested starting bid, and the negative to add back in auto so you don't pay twice.
2The "exact" campaigns aren't exact
Three campaigns named SKC-Exact-* actually contain phrase and broad match keywords — added during a rushed expansion and never audited. Result: the campaigns you believe are your precision layer are buying loosely-related traffic at exact-campaign bids. One "exact" campaign spent $322 last month on search terms that don't contain the keyword at all.
3Negative-keyword gaps — the $412/month leak
Six search terms with meaningful spend, zero orders, ever — each one a negative that should have been added months ago:
| Search term | Lifetime clicks | Lifetime spend | Orders | Fix |
|---|---|---|---|---|
| cheap [product category] | 214 | $387 | 0 | Negative phrase, all campaigns |
| [competitor brand] [product] | 166 | $298 | 0 | Negative exact (keep testing in a fenced campaign only) |
| [product] replacement parts | 143 | $211 | 0 | Negative phrase |
| commercial [product category] | 98 | $176 | 0 | Negative phrase (wrong buyer) |
Client version: every zero-converting term above a spend threshold, grouped into a paste-ready negative list per campaign.
4Three campaigns are bidding against each other
The hero SKU's main keyword is live in three campaigns simultaneously (auto, an old broad campaign, and the "exact" campaign). They take turns winning the same auctions, which inflates the effective CPC roughly 20% and makes every campaign's data unreadable. This is the single most common structural fault I find in self-built accounts.
5Top-of-search converts 2.4× — and gets no priority
Placement data shows top-of-search converting at 2.4× the rate of rest-of-search for the hero SKUs — and every top-of-search bid modifier in the account is at 0%. Meanwhile product-page placements eat 31% of spend at the account's worst conversion rate. The money is standing in the wrong queue.
6The budget rewards the losers
41% of daily budget sits in campaigns running above 60% ACOS, while the three campaigns running under 15% ACOS hit their budget caps by early afternoon — every single day. The best campaigns in this account are being throttled to fund the worst ones. Reallocating budget alone — before touching a single bid — moves this account's blended ACOS by an estimated 4–6 points.
7No structure, no names, no way to manage it
Campaign names like "New campaign — Aug 12" and "test2" aren't cosmetic problems: they're why nobody can tell what's working in under an hour, which is why nobody looks. The restructure ships a naming convention (SKU · goal · match type) that makes the account readable at a glance — the precondition for every weekly routine in the 90-day plan.
3 · The 90-day plan (first weeks shown)
| When | Action | Why first |
|---|---|---|
| Week 1 | Stop-loss: paste in the negative lists; cut product-page placement waste; shift budget from the >60% ACOS campaigns to the capped winners | Pure leak-plugging — no risk, immediate savings (~$1,900/mo of the $2,731 stops here) |
| Week 2 | Set top-of-search modifiers on hero campaigns; fence the competitor-targeting test into its own campaign | Reallocates spend toward the placement that converts 2.4× better |
| Weeks 3–4 | Harvest: launch exact campaigns for the 14 proven auto terms; add them as negatives in auto | Stops paying discovery prices for known demand |
| Weeks 5–8 | Restructure: dedupe the cannibalizing campaigns, rebuild to one-role-per-campaign, apply the naming convention | Makes the account readable and every future decision data-clean |
| Weeks 9–12 | Optimize: bid ladders on the new exact campaigns, weekly negative sweep routine, budget-reallocation cadence, KPI dashboard live | Locks the gains into a repeatable weekly system |
Client version: week-by-week with named campaigns, exact bid changes, and a 30-minute weekly routine your team can run without me.
4 · What to track (the KPI dashboard)
The client version ships a dashboard template. The short version — five numbers weekly, and where each one should trend for this account:
- Blended ACOS — 32.8% today; the plan targets the account's break-even band (≤27%) by week 6 and low-20s by week 12
- TACOS — the honest number; if ACOS falls while TACOS rises, we're just hiding spend
- Spend share in auto — from 38% toward ~15%: discovery budget, not the main engine
- Zero-order spend — dollars on terms with no lifetime orders; should trend to near-zero and stay there
- Winner throttle hours — how early your best campaigns hit budget; the goal is "they don't"