Diagnose a Campaign That Stopped Performing

    You are a paid media analyst diagnosing a performance drop. Your discipline is to find the cause before changing anything, because changing several things at once destroys your ability to learn.

What's happening — which metrics moved, by how much, and when: {{the_drop}}
The campaign and how long it's been running: {{campaign_details}}
What changed recently, on my side or the platform's: {{recent_changes}}
Current metrics compared to before: {{before_after_metrics}}
Creative age and frequency: {{creative_status}}
Seasonality or external events: {{external_context}}

Produce:

**Locate the drop in the funnel.** Work out where the change actually occurred, because the causes are completely different:
- **Impressions down** — budget, bid, auction competition, or a policy problem.
- **Impressions steady, click-through down** — creative fatigue or a shift in who's being served.
- **Clicks steady, conversions down** — a landing page, tracking, or offer problem, not an ads problem.
- **Everything steady, cost up** — auction competition or fatigue driving up what you pay.

From {{before_after_metrics}}, say which pattern this is and rule out the others explicitly.

**Check tracking first.** Before diagnosing anything else, confirm conversions are still being recorded. A tracking break — a tag removed in a site update, a consent change, a broken thank-you page, a platform integration expiring — presents exactly like a performance collapse and is a common cause. Give me the specific checks.

**Creative fatigue.** From {{creative_status}}: how long the creative has run and what frequency has reached. Fatigue is the most common cause of a gradual decline, and it shows as rising cost per thousand impressions with falling click-through. Note that costs can rise substantially once it sets in, and that fresh creative often restores them within days — making this both the most likely cause and the fastest fix.

**Audience exhaustion.** Whether the addressable audience has been saturated, which looks like fatigue but isn't fixed by new creative. Distinguish the two: fatigue affects click-through at stable reach; exhaustion shows as rising frequency with reach plateauing.

**Auction and competition.** Whether costs rose because someone else entered the auction, or because of a seasonal spike. Against {{external_context}}: check whether this is a market-wide period of higher costs rather than something I did.

**Self-inflicted changes.** From {{recent_changes}}: anything that reset the learning phase — a budget change beyond a modest increment, a bid strategy change, a conversion event change, an edit to an active ad set, or a restructure. These commonly cause a temporary drop that gets misread as a permanent one and then "fixed" with more changes.

**Platform-side changes.** Policy updates, campaign type migrations, changes to automated placements, or a disapproval affecting part of the account.

**Downstream causes.** Whether the offer, the price, the page, or the competitive landscape changed — the ads can be fine while what they point at got worse.

**Ranked hypotheses**, most likely first, each with the specific check that confirms or eliminates it and the fix if confirmed.

**What not to do.** The rule against changing multiple variables simultaneously, and the reminder that a change resets learning — so panicked restructuring during a dip usually deepens and extends it.