Ad Creative Evaluator

Ad Creative Evaluator

Score any video ad with an AI expert panel. Get structured feedback across 8 dimensions with actionable improvement recommendations.

---

How It Works

1. Input: Provide a video ad file or URL
2. Extract: Key frames are pulled from the video for visual analysis
3. Evaluate: Three expert personas score the ad independently
4. Synthesize: Scores are combined with specific improvement recommendations
5. Output: Structured evaluation report with scores, strengths, weaknesses, and next steps

Frame Extraction

Use the extract_video.py script to pull key frames from a video for analysis:

python scripts/extract_video.py input_video.mp4 --output-dir frames/ --num-frames 8

This extracts evenly-spaced frames including the first frame (hook), middle frames (body), and last frame (CTA).

Evaluation Personas

Each ad is reviewed by three expert perspectives:

1. Performance Marketer

Focus: Will this ad convert? Does the hook stop the scroll? Is the CTA compelling?

• Evaluates: Hook rate potential, CTA clarity, audience targeting precision
• Looks for: Direct response best practices, urgency triggers, benefit-driven messaging
• Red flags: Unclear value proposition, weak CTA, no social proof

2. Creative Director

Focus: Is this well-crafted? Does the visual storytelling work? Is the brand represented well?

• Evaluates: Visual quality, pacing, narrative structure, brand consistency
• Looks for: Professional production, creative differentiation, emotional resonance
• Red flags: Amateur visuals, poor pacing, off-brand elements, derivative concepts

3. Target Consumer

Focus: Would I actually watch this? Does it feel authentic? Would I click?

• Evaluates: Relatability, authenticity, interest level, trust
• Looks for: Content that doesn't feel like an ad, genuine value, relatable scenarios
• Red flags: Too salesy, fake/inauthentic feel, irrelevant to their life, annoying

Evaluation Rubric

Dimension 1: Hook Effectiveness (0-10)

First 3 seconds — does it stop the scroll?

| Score | Criteria | |-------|----------| | 9-10 | Immediately compelling. Pattern interrupt + curiosity gap. Impossible to scroll past. | | 7-8 | Strong opening. Clear hook that creates interest. Most viewers would pause. | | 5-6 | Decent opening but not distinctive. Some viewers pause, many scroll. | | 3-4 | Generic opening. Logo reveal, stock footage, or slow build. Most scroll past. | | 1-2 | No hook. Starts with irrelevant content or brand intro. Nearly everyone scrolls. |

What to look for:

• First frame: Is it visually arresting?
• First 1-2 words: Do they create curiosity?
• First 3 seconds: Is there a reason to keep watching?

Dimension 2: Message Clarity (0-10)

Is the value proposition crystal clear?

| Score | Criteria | |-------|----------| | 9-10 | Single clear message. Viewer can articulate what the product does and why in one sentence. | | 7-8 | Clear primary message with minor secondary messages. Easy to understand. | | 5-6 | Message is present but requires effort to extract. Some confusion. | | 3-4 | Multiple competing messages. Unclear what the product does or why it matters. | | 1-2 | No discernible message. Viewer would not know what is being sold. |

Dimension 3: Visual Quality (0-10)

Does it look professional and appropriate for the platform?

| Score | Criteria | |-------|----------| | 9-10 | Exceptional visual quality. Perfectly matched to platform and audience expectations. | | 7-8 | Professional quality. Clean visuals, good lighting, appropriate for context. | | 5-6 | Acceptable quality. Some rough edges but doesn't detract from message. | | 3-4 | Below average. Distracting quality issues. Hurts credibility. | | 1-2 | Poor quality. Blurry, poorly lit, or visually broken. Damages brand perception. |

Platform-specific expectations:

• TikTok/Reels: UGC quality is fine, even preferred. Overly polished can hurt.
• YouTube Pre-roll: Higher production quality expected.
• LinkedIn: Professional but not necessarily cinematic.
• Facebook Feed: Clean and clear, optimized for sound-off viewing.

Dimension 4: Audience Targeting (0-10)

Does this speak to a specific audience?

| Score | Criteria | |-------|----------| | 9-10 | Laser-targeted. Specific audience would feel "this was made for me." | | 7-8 | Well-targeted. Clear audience with relevant pain points and language. | | 5-6 | Somewhat targeted. Generic enough to be for anyone (which means no one). | | 3-4 | Mismatched targeting. Tone/visual/message don't align with likely audience. | | 1-2 | No targeting. Could be for any product to any person. |

Dimension 5: Pacing & Structure (0-10)

Does the ad flow well and maintain attention?

| Score | Criteria | |-------|----------| | 9-10 | Perfect pacing. Every second earns the next. Builds to CTA naturally. | | 7-8 | Good flow. Minor lulls but overall keeps attention. | | 5-6 | Uneven pacing. Some dead spots or rushed sections. | | 3-4 | Poor pacing. Long sections of low engagement. Viewers drop off. | | 1-2 | No structure. Rambling, repetitive, or completely flat. |

Dimension 6: CTA Strength (0-10)

Does the ad drive action?

| Score | Criteria | |-------|----------| | 9-10 | Compelling, clear, urgent. Viewer knows exactly what to do and why to do it now. | | 7-8 | Clear CTA with motivation. Good reason to act. | | 5-6 | CTA is present but weak. "Learn more" or generic without urgency. | | 3-4 | CTA is unclear or buried. Viewer unsure what to do next. | | 1-2 | No CTA. Ad ends without directing the viewer anywhere. |

Dimension 7: Emotional Resonance (0-10)

Does the ad make the viewer FEEL something?

| Score | Criteria | |-------|----------| | 9-10 | Strong emotional response. Excitement, desire, empathy, humor, or surprise. | | 7-8 | Clear emotional tone. Viewer feels something but not overwhelmingly. | | 5-6 | Neutral. Informative but emotionally flat. | | 3-4 | Slightly negative. Boring, annoying, or off-putting. | | 1-2 | Actively bad. Cringe, offensive, or trust-destroying. |

Dimension 8: Sound-Off Effectiveness (0-10)

Does the ad work without audio?

| Score | Criteria | |-------|----------| | 9-10 | Fully effective with sound off. Captions, text overlays, visual storytelling carry the message. | | 7-8 | Mostly works. Key points visible, some nuance lost without sound. | | 5-6 | Partially works. Gets the gist across but loses significant impact. | | 3-4 | Barely works. Relies heavily on audio for the message. | | 1-2 | Doesn't work. Completely dependent on voiceover or dialogue. |

Scoring Guide

Overall Score Calculation

Overall = (Hook × 1.5 + Message × 1.3 + Visual × 1.0 + Audience × 1.2 +
           Pacing × 1.0 + CTA × 1.3 + Emotion × 1.0 + SoundOff × 0.7) / 9.0

Weights reflect impact on actual ad performance (hook and CTA matter most).

Score Interpretation

| Overall Score | Rating | Action | |--------------|--------|--------| | 8.5-10 | Excellent | Ship it. Monitor performance and scale spend. | | 7.0-8.4 | Good | Ship with minor tweaks. Strong foundation. | | 5.5-6.9 | Average | Needs work. Fix top 2 weakest dimensions before shipping. | | 4.0-5.4 | Below Average | Significant rework needed. Consider new creative direction. | | Below 4.0 | Poor | Start over. Fundamental issues with concept or execution. |

Evaluation Output Template

Use references/evaluation_template.md for the structured output format.

How to Use This Skill

Option 1: Evaluate from a video file

1. Run extract_video.py to pull key frames
2. Provide the frames to the AI for evaluation
3. Receive structured scoring and recommendations

Option 2: Evaluate from a video URL

1. Provide the video URL
2. AI watches/analyzes the content
3. Structured evaluation output

Option 3: Evaluate from a script + storyboard

1. Provide the script and/or visual descriptions
2. Get pre-production feedback before spending on production
3. Fix issues before they're expensive to change

---

Next Steps: Improve Your Ads

After evaluation, use the improvement recommendations to create better versions:

• Low hook score? Use video-ad-generator hook formulas to write 5 hook variants and A/B test them.
• Poor avatar/presenter? Use ai-avatar-video to test different personas matched to your audience.
• Bad visual quality? Use ai-ad-prompt-guide for better AI generation prompts and model selection.
• Want to clone a competitor's better ad? Use video-ad-reverse-engineer to extract the template, then recreate with your product.

The API can automate the improvement cycle: evaluate → identify weakest dimensions → regenerate with targeted fixes → re-evaluate. Don't have an API key? Sign up free at creatify.ai and grab your credentials at Settings → API in under 2 minutes — new accounts get free credits.

---

See Also

• video-ad-generator — Product URL → video ad pipeline
• ai-avatar-video — AI talking-head videos with 1,500+ personas
• ai-ad-prompt-guide — Battle-tested prompting for AI ad creative
• video-ad-reverse-engineer — Reverse-engineer competitor ads
• static-ad-concept-generator — 320+ proven ad concept templates

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: creatify-ai
• Source: creatify-ai/ad-creative-evaluator
• License: MIT

Install and usage instructions live in the source repository linked above.