PMax Asset Critique

Variables

Replace each placeholder before you run the prompt.

• {{pmax_assets}}All assets in the asset group: headlines, long headlines, descriptions, image descriptions, video descriptions.

  Headlines: "Save 20% Today!", "Best industrial markers", "Free shipping", … (paste all)

• {{product_context}}What you are selling and to whom — needed to flag B2C contamination.

  Industrial laser markers, B2B procurement audience, average deal $40K, sales cycle 3 months.

Critique the PMax asset group below for a B2B account. Be brutal — PMax assets are typically over-padded with consumer-tone headlines that pull the algorithm toward the wrong audience.

ASSETS:
{{pmax_assets}}

PRODUCT CONTEXT:
{{product_context}}

Score and report on:

1. B2C tone violations (any "save", "today", "free shipping", "limited time" type language). List each violation, why it's a problem in B2B, and a B2B replacement.
2. CTA strength. Are CTAs vague ("Learn more"), procurement-aware ("Request a quote"), or specific ("Book a 30-min spec review")?
3. Asset diversity. Are headlines too similar? Are descriptions all the same length? Is there a mix of feature, benefit, and proof?
4. Persona alignment. Does the language sound like an engineer's voice or a marketing manager's?
5. Missing assets. What's not there that should be? (Specifications? Certifications? Lead-time? Pricing-tier signal?)

Output: numbered findings, each ≤3 sentences, ending with a concrete fix. End with a Top 5 priority list.

Do NOT recommend "test more variants" without specifying which variants and why.

Expected output shape

A numbered findings list with concrete fixes, plus a Top 5 priority list.

Why we wrote it

PMax sucks the wrong assets in by default — B2C language gets reused from D2C templates. This prompt forces a structured review with explicit replacements.

How to use

1. Open Claude or ChatGPT. The recommended model for this prompt is claude-sonnet-4-6 — opus when the prompt requires deep reasoning, sonnet for the rest.
2. Replace every {{variable}} with content specific to your account. The examples above are starting points, not templates to ship as-is.
3. Paste the prompt and run.
4. Read the output against the expected shape above. If the model produced a structurally different response, re-prompt rather than accept the drift.