Diagnose Underperformer

---
name: diagnose-underperformer
description: Diagnose why a specific Google Ads or Meta Ads campaign / ad set / ad is underperforming and prescribe the next concrete action. Use when the user says something like "why is X not converting?", "this campaign is wasting money", or "why are impressions so low?". Read-only — no changes are made.
---
## Goal

Pull the right report, walk a structured decision tree, and hand the user a ranked list of likely causes with the **specific next op to call** to fix it. No writes — diagnosis only.

## Inputs to confirm (batch)

1. Account name (substring) — required.
2. Target — campaign / ad set / ad name or ID — required.
3. Date range — default last 14 days; if the entity is < 14 days old, use its lifetime.
4. Comparison window (optional) — previous 14 days, for trend.
5. Primary KPI — conversions / ROAS / CPA / CPL — required if the entity has a conversion goal.

## Workflow

1. **Discover** ops. For Google Ads: `gads_select_accounts`, `gads_list_report_fields`, `gads_get_report`, `gads_get_account_history`. For Meta: `meta_ads_select_accounts`, `meta_ads_list_report_fields`, `meta_ads_get_report`, `meta_ads_get_account_history`, `meta_ads_get_adset_settings`, `meta_ads_get_campaign_settings`.
2. **Inspect** with `get_operation_inputs`. **Always call `*_list_report_fields` before `*_get_report`** — never guess metric/dimension names.
3. **Resolve account** with substring match.
4. **Pull the report** at the right granularity (campaign / ad set / ad). Required metrics:
   - Spend, impressions, clicks, CTR, CPC, conversions, conv rate, CPA / ROAS, frequency (Meta), search impression share + lost-to-budget + lost-to-rank (Google Search).
5. **Pull settings/history** to check for recent edits that could have triggered the dip:
   - Google: `gads_get_account_history` filtered to this campaign — look for budget changes, bid strategy switches, status flips, asset changes.
   - Meta: `meta_ads_get_campaign_settings` / `meta_ads_get_adset_settings` for current state, `meta_ads_get_account_history` for recent changes.
6. **Walk the decision tree below** and rank the matching diagnoses by likelihood + impact.
7. **Present**: top 3 diagnoses, evidence from data, and the exact next op to call (or the matching slash command). Do not run any write op.

## Decision tree

### Branch A — Low impressions / low spend (entity not delivering)

| Signal | Diagnosis | Next action |
|---|---|---|
| Google: search impression share lost to budget > 20% | Budget-capped | Raise daily budget or use `/markifact:negative-keyword-sweep` to free spend |
| Google: search impression share lost to rank > 30% | Bid / Quality Score too low | Raise CPC bid (manual) or improve ad relevance (Quality Score components) |
| Google: campaign status `LIMITED` (account history) | Eligibility / policy / payment / learning | Inspect `gads_get_account_history` for the cause; route based on subtype |
| Meta: ad set status `IN_REVIEW` or `REJECTED` | Policy | Check creative against Meta policies; replace creative |
| Meta: estimated audience size < 500K and CBO is on | Audience too narrow + CBO starved it | Widen audience or move to ABO with fixed budget |
| Meta: > 4 ad sets in same campaign with CBO | Auction overlap, CBO concentrating spend on one | Consolidate ad sets or split into separate campaigns |
| Both: budget set but daily spend << budget | Bid too low for the auction OR audience tiny OR creative not eligible | Inspect bid + audience size + creative status |

### Branch B — Spending normally, low CTR

| Signal | Diagnosis | Next action |
|---|---|---|
| Meta: frequency > 3.5 + CTR dropping WoW | Creative fatigue | `/markifact:rotate-creative` |
| Google: RSA ad strength `Poor` / `Average` | Weak asset variety | Add more headline variety, re-create RSA |
| CTR way below industry baseline (Search < 3%, Meta feed < 0.8%) and frequency low | Wrong audience or weak hook | Tighten targeting; try a new creative angle |
| Branded search CTR < 5% | Competitor bidding on brand or weak ad copy | Check `gads_ads_transparency` for competitors; strengthen brand copy / pin brand headline |

### Branch C — Good CTR, low conversion rate

| Signal | Diagnosis | Next action |
|---|---|---|
| Bounce rate high in GA4 / time on page short | Landing page mismatch | Send fresh LP variant; consider `/markifact:edit-meta-creative` to swap URL |
| Conversion event firing but CVR < 0.5% | LP friction / wrong offer / wrong audience intent | Audit funnel; consider gating with a softer CTA |
| Conversion not firing at all in last 7 days | Pixel / tag broken | Meta: `meta_ads_check_capi_health` and `meta_ads_get_pixel_stats`; Google: check conversion action import from GA4 |
| Mobile CVR << desktop CVR | Mobile LP problem | Fix mobile LP; consider device bid adjustment |

### Branch D — Good CVR, low overall conversions

| Signal | Diagnosis | Next action |
|---|---|---|
| Spend capped (see Branch A signals) | Budget-bound winner | Raise budget incrementally (≤ 20%/day on Smart Bidding) |
| Audience small + saturated (Meta freq > 4) | Audience exhausted | Build lookalike from converters; expand interests |
| Strong perf only on weekends | Schedule misalignment | Adjust ad schedule / bid by hour |

### Branch E — Sudden performance drop

| Signal | Diagnosis | Next action |
|---|---|---|
| Account history shows a bid strategy switch in the last 14 days | Bid strategy reset learning | Either revert or wait 14 days; warn about future switches |
| Budget cut > 30% recently | Underfunded learning | Restore budget or accept slower learning |
| Creative replaced recently with major change | Soft learning reset | Wait 7 days before judging |
| Seasonal / industry-wide | Compare to category trend | Note and move on; not a config issue |

## Platform-specific rules the model MUST respect

- **`list_report_fields` first.** Field names differ per platform; never guess.
- **Account-name substring** match for `select_accounts`.
- **Frequency is a Meta concept** — not directly comparable on Google.
- **Search impression share** is a Google Search-specific metric; PMax has its own (search-categories report).
- **Don't propose changes that need to wait.** If the entity is < 7 days old or made a major change in the last 7 days, say "still in learning, don't touch yet."
- **Read-only.** This command never calls a write op. Recommend the right write command instead.

## Output to user

```
Top diagnoses for <entity>:

1. <Diagnosis> — confidence: high/medium
   Evidence: <numbers>
   Next: run /markifact:<command> OR call <op_id>

2. ...

3. ...

Other observations:
- ...
```

End with: *"Want me to do any of these? I won't make changes without your go-ahead."*