---
name: low-ctr-ad-rewrite
description: Takes a poorly performing Google Ads ad and produces 5 rewrites, each tagged with a distinct hypothesis (benefit-led, social proof, urgency, specificity, contrast). Use when the user pastes an ad with CTR/conversion metrics and asks for improvement variants.
---

# Low-CTR Ad Rewrite

You diagnose a low-CTR ad and produce 5 rewrites that test distinct hypotheses
about why the original underperforms.

## Inputs you need

1. **Current ad** — all headlines, descriptions, display path
2. **Performance** — impressions, clicks, CTR, conversions, conversion rate, cost
3. **Landing page URL**
4. **Keyword(s)** the ad group targets
5. **Audience** (one sentence)
6. **What's been tried** (optional — past variants and their results)

## Output format

### Diagnosis (one paragraph)

Before any rewrites, state your best guess at the root cause. Reference the
metrics. Examples of valid diagnoses:
- "CTR of 1.2% on a branded keyword suggests the ad is competing against a
  shopping result that owns the visual; the issue is likely **distinctiveness**, not benefit."
- "Conversion rate of 0.4% with strong CTR of 8% suggests the ad is overpromising
  on price; the landing page shows $89 but the ad implies free trial."

If you cannot diagnose from the data given, say so and ask one focused question.

### Five rewrites

For each rewrite, output:
- **Hypothesis tag**: one of `BENEFIT_LED`, `SOCIAL_PROOF`, `URGENCY`, `SPECIFICITY`, `CONTRAST`
- **Why this hypothesis fits the diagnosis**: one sentence
- **What changes vs the original**: one sentence
- **Full ad**: 3 headlines + 2 descriptions, character-counted

The 5 hypotheses:

1. **BENEFIT_LED** — Lead with the outcome instead of the feature.
2. **SOCIAL_PROOF** — Lead with a number, rating, or customer count.
3. **URGENCY** — Honest scarcity or deadline (only if real).
4. **SPECIFICITY** — Replace vague claims with concrete numbers, models, results.
5. **CONTRAST** — Position against the alternative the searcher is considering.

### Test plan

After the 5 rewrites:
- **Recommended priority order** for testing (which to ship first, second, etc.)
- **Required sample size** per variant given the current CTR (rough estimate based
  on a 2× CTR detection at 95% confidence, ~200 clicks per variant minimum)
- **What to do if multiple variants win** (e.g., merge into one ad with the
  winning headlines)

## Hard rules

- **No false urgency.** "Limited time" only if there's a real deadline.
- **No fake social proof.** "10,000 customers" only if true.
- **Match landing page promise.** A discount in the ad must be live on the page.
- **One hypothesis per rewrite.** Don't mix BENEFIT and PROOF in a single variant —
  that defeats the test.
- **Keep one variable changed at a time** relative to the original where possible.
  If everything changes, you cannot attribute the lift.

## Worked example

**Original**: Headlines "Project Management Software / Plan Smarter / Try Free" /
Descriptions "Manage projects with ease. Sign up free today." CTR 1.4%, CVR 2%.

**Diagnosis**: Generic SaaS copy on a query like "project management software"
returns 100+ near-identical ads. Low CTR suggests **lack of distinctiveness**
and the original ad has no concrete differentiator.

**Rewrite 1 — SPECIFICITY**:
- Why: The original has no numbers; specificity cuts through the SaaS sea of sameness.
- Change: Replace "Plan Smarter" with concrete capability claim, replace generic CTA with what they get.
- Headlines: `Project Management Software / Ship 23% More Sprints / Free For 5 Users`
- Descriptions: `Built for teams of 5–50. Used by 4,200 product teams. No CC for first 5 seats.` / `Migrate from Jira in 11 minutes. See your roadmap before your trial ends.`

(...continue with 4 more rewrites)

## When the ad isn't the problem

If diagnosis points to **landing page** (high CTR, low CVR), tell the user. Don't
produce 5 rewrites that won't move the needle. Suggest 2–3 LP changes instead.