The Prompt
You are a senior B2B paid media analyst auditing video and influencer ad creative performance. You manage paid programs across LinkedIn (video view campaigns), Meta (traffic and conversion campaigns), and connected TV. You know that creative decisions made on top-funnel metrics alone (impressions, view rate) destroy budget — and that creative needs to be tied to down-funnel impact (MQLs, pipeline) before scaling or killing.
INPUTS
I will paste creative-level performance data below. The data should include creative-level metrics from the ad platforms (LinkedIn, Meta) and down-funnel MQL or lead data merged in. Required fields per row: creative name / influencer, channel, impressions, view rate, 50% view rate (LinkedIn) or video play rate (Meta), CPL, leads, MQLs.
{PASTE_CREATIVE_DATA_HERE}
{OPTIONAL_PASTE_CONTEXT_HERE}
(Examples: "Influencer X has been running 3 weeks at $5K/week," "Video B was a recent re-edit of Video A." Leave blank if no context to add.)
WHAT I NEED FROM YOU
Audit each creative variant against the four scoring dimensions below. Produce the output in this exact order:
1. Creative Leaderboard
A ranked table of all creative variants, scored on each of the four dimensions:
- View Rate (top-funnel attention)
- 50% View Rate (LinkedIn retarget pool eligibility) or video play rate (Meta proxy)
- CPL (direct efficiency)
- MQL pull-through (down-funnel quality — MQLs / leads)
Rank by composite score: top-funnel signals weight 30%, CPL weights 30%, MQL pull-through weights 40%. Down-funnel quality is the deciding factor.
2. Scaling Recommendations
For each top-quartile creative: recommend scaling action (increase budget, add new placements, expand audiences, build retarget pool from 50% viewers). Be specific about the dollar amount or percentage increase.
3. Kill List
Creatives in the bottom quartile with at least 2 weeks of run-time. State the creative, the underperforming metrics, and recommend either: kill, recut, or move to lower-funnel as retargeting only.
4. Pattern Observations
2-3 observations about what's working across the top-performers and what's failing across the bottom. Pattern-level insights, not just list rehashing.
JUDGMENT RULES
- Do not kill a creative based on under 2 weeks of run-time. Short windows produce noise. Flag short-run creatives as "insufficient data" rather than including in the kill list.
- View rate alone is not a kill signal. A creative with a low view rate that produces strong MQL pull-through is doing its job — the top-funnel metric is just optimization noise.
- CPL on Meta should always be evaluated alongside MQL pull-through. Meta is great at producing cheap leads of unclear quality. A $20 CPL with 5% MQL pull-through is worse than a $60 CPL with 30% MQL pull-through.
- 50% View Rate on LinkedIn matters as a retargeting input even if the creative isn't a leader. A creative with a high 50% view rate is feeding the retargeting pool, which has downstream value.
- If MQL data is missing from the input, say so explicitly and rank using top-funnel and CPL only. Flag that the audit is incomplete without MQL data. Do not infer MQL counts from CPL alone.
OUTPUT FORMAT
Return as {OUTPUT_FORMAT}.
If "markdown": leaderboard table, then scaling and kill sections, then patterns.
If "html": tabbed artifact structure (Leaderboard / Scaling / Kill / Patterns).
Begin.
How to Use It
This prompt solves a specific problem: creative teams and paid media managers evaluate creative on different metrics, and neither set alone tells the right story. Creative teams look at view rate; paid teams look at CPL. Both miss the point. View rate without down-funnel conversion tells you attention without impact; CPL without MQL pull-through tells you cost without quality. The four-dimension scoring framework in this prompt forces both sides of that equation — with MQL pull-through weighted highest at 40% because it's the metric closest to revenue.
Claude handles the multi-dimensional scoring and creative ranking well. GPT-4 class models also work, but they sometimes underweight MQL pull-through relative to top-funnel metrics when the data shows tension between them. The 40% weight on MQL pull-through is intentional and should be preserved. If you need to adapt the weights for a specific use case (e.g., top-funnel brand awareness campaigns where MQL pull-through isn't the goal), note that explicitly in the context input.
The 2-week minimum run-time rule is the most important judgment call in the prompt. Running this analysis on creative with 3–5 days of data produces noise, not signal. If you have creative that launched last week, exclude it from the current audit and revisit in 10 days. The model flags short-run creatives as "insufficient data" rather than including them in the kill list — that instruction is intentional.
Example Output
Example output coming soon — currently running this prompt against live data and will publish the redacted output once it's ready.
Common Failure Modes
Variations
Two variations of this prompt are worth knowing.
Static Image Creative Audit
Adapted for static ad creative (LinkedIn sponsored posts, Google Display) using CTR, engagement rate, CPL, and MQL pull-through instead of video-specific metrics. Once you've identified winning static creative, use the Ad Copy Variant Generator to iterate on the copy angles that are working.
[PROMPT GOES HERE]
Monthly Creative Portfolio Review
Same logic but scoped to a full monthly review across all creative in flight, suitable for a monthly creative retrospective with the creative team. Includes quarter-over-quarter trend analysis across the creative portfolio.
[PROMPT GOES HERE]
Get one new prompt every Monday.
Plus the system behind it. Free. Built for in-house demand gen managers at B2B SaaS companies.
Subscribe free →Frequently Asked Questions
Does this work with ChatGPT or only Claude?
Both work. Claude is the preferred choice because it's more reliable about applying the 40% MQL pull-through weighting consistently, even when top-funnel metrics tell a different story. GPT-4 sometimes reverts to a simpler CPL ranking when the composite scoring logic creates unexpected orderings. For creative decisions that influence budget allocation, the composite score is the right signal — use Claude if you need it honored precisely.
What if I don't have MQL data merged into my creative-level report?
The prompt will flag the audit as incomplete and rank on the remaining three dimensions (view rate, 50% view rate, CPL) only. That's still useful for top-funnel creative evaluation, but you're missing the most important signal. Setting up MQL pull-through reporting at the creative level — even as a manual merge from Salesforce — is worth the setup time because it changes almost every creative decision you make.
How do I interpret the 50% view rate metric for LinkedIn retargeting?
A high 50% view rate means that creative is building your LinkedIn retargeting pool — the audience of people who watched at least half the video, which is the signal LinkedIn uses for video-based retargeting audiences. Even if a creative isn't a CPL leader, high 50% view rate has downstream pipeline value because it feeds lower-funnel retargeting. The prompt accounts for this — a creative with a high 50% view rate won't appear on the kill list even if other metrics are mid-tier.
Should I audit LinkedIn and Meta creative together or separately?
Separately. LinkedIn and Meta have different attention economics, audience compositions, and conversion pathways — a creative that performs well on LinkedIn will often underperform on Meta and vice versa. Run the audit per platform, then look at cross-platform patterns as a separate observation. The Pattern Observations section of the output will surface these if you tell the model which creative ran on multiple platforms.