Performance Triage (Operator Playbook)
Use this when a campaign is underspending, too expensive, or not converting.
Quick principle: diagnose delivery first, then click quality (CTR/CPC), then conversion quality (CVR/ROAS).
First 30 minutes: triage summary
| Symptom | Most common root cause | First actions |
|---|---|---|
| Underspending | bids too low, targeting too narrow, creative rejected, eligibility issues | check eligibility/approvals, broaden targeting, raise bids in small steps |
| High CPC / CPM | low relevancy, too competitive targeting, weak listing quality | tighten to higher-intent targets, improve PDP/offer, isolate placement intent |
| Low CTR | creative/offer not resonating, wrong audience, weak placement fit | refresh hook/message, change targeting type, simplify offer |
| Good CTR but low CVR | PDP/price/stock/buy box, intent mismatch, landing mismatch | fix retail readiness, refine keywords, align landing page |
| Good sales but weak ROAS | bids too high, overbroad terms, mix of placements hiding loss | add query controls (negatives/match types), split campaigns, lower bids |
Foundation checks (do these before you touch bids)
These issues make everything look broken: - Item in-stock and winning buy box (critical for Sponsored Products placements) - Creative approved in Creative Hub / no policy issues - Item page quality (title, images, reviews) is not obviously weak - Landing page matches the promise (Shelf / Brand Shop merchandising)
Reference: - Tactics & Ops
Time horizon: what to change when
First 30 minutes (same-day fixes)
Do: - Fix approvals/eligibility, stock/buy box, landing mismatch. - Make one change at a time per campaign (or isolate in a duplicate).
Avoid: - Sweeping bid changes across everything. - Rebuilding targeting before you validate delivery.
First 24 hours
Do: - Pull early placement/targeting signals (what is getting impressions?). - Decide if you’re in a learning phase (launch/seasonal ramp) or efficiency phase (always-on).
Week 1
Do: - Split campaigns so you can turn knobs cleanly (Search vs PDP vs Retargeting). - Build query controls (phrase/exact, negatives) from early search terms.
Decision trees by symptom
A) Underspend / not delivering
Goal: get eligible, broaden inventory, then pay enough to win auctions.
1) Is the campaign eligible + approved? - If no: fix policies/creative approvals first. - If yes: continue.
2) Are advertised items in-stock and winning buy box? - If no: pause ads for that item or switch the advertised SKU set. - If yes: continue.
3) Is targeting too narrow? - Sponsored Products: ensure you have enough keywords/targets; avoid only ultra-long-tail. - Display: expand audiences (persona ’ ROS ’ contextual) before raising bids.
4) Are bids high enough for the placement intent? - Raise bids in controlled steps (e.g., 10–20%) and observe impression movement. - If you can’t win without breaking economics: change placement intent (lower-cost inventory) rather than forcing it.
5) Is daily budget inadvertently capping you? - If budgets run out early: increase daily budget or change delivery strategy.
B) High CPC (search) or high CPM (display)
Goal: reduce waste by increasing relevancy and controlling where you compete.
1) Did CPC/CPM rise because competition rose (seasonality/event)? - If yes: expect higher costs; shift to more efficient inventory or tighten to higher intent.
2) Are you mixing different intents in one campaign? - If yes: split by intent: - Search (high intent) - PDP conquest (mid intent) - Retargeting (high relevance)
3) Is relevancy weak? - Improve PDP readiness (title clarity, imagery, reviews, price competitiveness). - Tighten targeting to fewer, more relevant queries/audiences.
4) Are you paying for low-quality clicks? - Add negatives; move from broad ’ phrase/exact; remove irrelevant keyword themes.
C) Low CTR
Goal: either change the message or change who sees the message.
1) Is the creative promise clear in the first second? - For video: stronger first frame, on-screen text, single value prop. - For display/headlines: simplify; one offer, one reason.
2) Is targeting aligned to the message? - Prospecting creative should be benefit-led (not SKU-detail-heavy). - Retargeting creative should be offer-led (price/urgency) and SKU-specific.
3) Are you in the wrong placement for the asset? - If CTR is weak on broad placements, test a higher-intent placement (search) or narrower audience.
D) Good CTR but low CVR
Goal: fix retail readiness or intent mismatch.
1) Check buy box + in-stock + shipping promise. - If unstable, fix retail before spending more.
2) Check price vs competitor set. - If you’re expensive with no clear advantage, CVR will stay weak.
3) Check PDP quality. - Title clarity, images, variants, reviews momentum.
4) Check intent mismatch. - Search terms: are you buying research queries for a conversion campaign? - PDP targets: are you conquesting without a clear reason to switch?
E) Good sales but weak ROAS
Goal: keep what converts, cut what steals budget.
1) Is ROAS weak because you’re scaling too fast? - If yes: pull back to the best-performing targets until you can expand methodically.
2) Do you have query/placement controls? - If no: add negatives; move to phrase/exact; split campaign intents.
3) Are you mixing new-to-brand awareness with efficiency reporting? - If yes: report separately (awareness campaigns should not be judged on same ROAS bar).
Funnel-stage context (what ’good’ looks like)
- Awareness: expect weaker ROAS; optimize toward reach/CTR and creative learnings.
- Conversion: ROAS and CVR are the primary truth; keyword/placement controls matter most.
- Loyalty: audience segmentation matters more than keywords; focus on repeat and retention.
Reference: - Strategic Framework - Benchmarks & Guardrails