Thursday, 24 September 2026
D Data-Driven Growth Studio
Digital Marketing

AI vs. Human Ads: Thread & Loom’s 2025 ROAS

Listen to this article · 10 min listen

Key Takeaways

  • AI-generated content can significantly reduce creative production costs and accelerate campaign launches, as demonstrated by a 40% reduction in creative spend for our target campaign.
  • Human-crafted ad creatives consistently achieve higher engagement rates and lower cost-per-conversion, with this campaign showing a 15% better ROAS for human-led ads despite higher initial production costs.
  • A hybrid approach, combining AI for initial content generation and human oversight for refinement, delivers the strongest overall ad effectiveness, improving CTR by 1.2 percentage points compared to purely AI or human efforts.
  • Effective AI integration requires clear, detailed prompts and iterative feedback loops to align outputs with brand voice and campaign objectives, preventing generic or off-brand messaging.
  • Continuous A/B testing between AI and human creative variants is essential for identifying optimal performance drivers and informing future content strategy.

The integration of artificial intelligence into marketing creative workflows has sparked considerable debate, particularly regarding its impact on ad effectiveness. We recently executed a campaign comparing the performance metrics of AI-generated content against human-crafted ads for a new direct-to-consumer (DTC) apparel brand. This teardown dissects the strategies, outcomes, and critical learnings from a Q4 2025 launch campaign with a total budget of $250,000 across paid social and search.

Initial Creative Generation
AI generates content (Midjourney, Copy.ai) in 20 hours for $15,000.
Human Crafting
Designers/copywriters create ads over 120 hours for $25,000.
Campaign Execution
Run parallel AI and human ad tracks for 8 weeks.
Performance Analysis
Compare ROAS, CTR, and CPC across both tracks.
Iterative Optimization
A/B test and refine based on human ads’ 15% better ROAS.

Campaign Overview: Launching “Thread & Loom”

Our client, “Thread & Loom,” aimed to introduce its line of sustainably sourced, minimalist clothing to a US audience aged 25-45, primarily through Instagram, Facebook, and Google Search. The core objective was to drive initial sales and build brand awareness. We structured the campaign into two parallel tracks: one using AI-generated ad creatives and copy, and the other employing traditional human designers and copywriters. The campaign ran for eight weeks, from October 1st to November 25th, 2025. We allocated 60% of the budget to paid social (Instagram and Facebook) and 40% to Google Search Ads. The targeting focused on individuals interested in sustainable fashion, ethical consumption, and minimalist aesthetics, with income tiers exceeding $75,000 annually. Geographically, we concentrated on major metropolitan areas like Atlanta, Austin, and Portland, where these demographics are highly concentrated.

Strategy and Creative Approach

The overarching strategy was to test the hypothesis that AI could produce cost-effective, scalable ad creatives without sacrificing performance significantly. For the AI track, we used a combination of generative AI platforms: Midjourney for image generation and Copy.ai for ad copy and headlines. Our prompts for Midjourney included specific aesthetic guidelines: “minimalist, natural light, diverse models, sustainable fabric textures, urban background.” For Copy.ai, we fed it brand guidelines, product descriptions, and target audience profiles, instructing it to generate short-form ad copy emphasizing sustainability and comfort. The human track involved a small team of two graphic designers and one copywriter. They developed creatives based on the same brand guidelines and product information, focusing on original photography and hand-crafted copy. This team spent approximately 120 hours developing the initial creative assets over a three-week period. In contrast, the AI creative generation, including prompt engineering and selection, took roughly 20 hours.

Targeting Parameters

Both campaign tracks shared identical targeting parameters to ensure a fair comparison. On Meta platforms (Facebook and Instagram), we used interest-based targeting for “sustainable fashion,” “ethical consumerism,” “minimalist lifestyle,” and “eco-friendly products.” We also layered in demographic targeting for age (25-45), gender (all), and income. For Google Search Ads, we targeted keywords such as “sustainable clothing brands,” “ethical apparel,” “organic cotton t-shirts,” and “minimalist wardrobe.” We implemented negative keywords to filter out irrelevant searches, like “fast fashion” or “cheap clothes.”

Performance Metrics: A Head-to-Head Comparison

The campaign metrics provided a clear, if nuanced, picture of AI versus human performance. All data reflects the full eight-week campaign duration.

Overall Campaign Budget: $250,000

Campaign Duration: 8 weeks (Oct 1 – Nov 25, 2025)

Creative Production Costs

This was one of the most stark differences. The AI-generated content track incurred approximately $15,000 in creative production costs (platform subscriptions, prompt engineering time). The human-crafted ads track had creative production costs of $25,000 (designer/copywriter salaries, photography fees). This represents a 40% reduction in creative spend for the AI approach.

Paid Social Performance (Meta Platforms)

Total Social Ad Spend: $150,000 ($75,000 per track)

Metric AI Track Human Track
Impressions 12,500,000 11,800,000
Clicks 187,500 212,400
Click-Through Rate (CTR) 1.50% 1.80%
Conversions (Purchases) 2,800 3,600
Cost Per Conversion (CPC) $26.79 $20.83
Return on Ad Spend (ROAS) 2.1x 2.5x

On Meta platforms, the human-crafted ads outperformed AI in terms of engagement and conversion efficiency. The CTR for human ads was 1.80%, notably higher than the 1.50% for AI-generated ads. This translated directly to more conversions (3,600 vs. 2,800) and a lower Cost Per Conversion (CPC) of $20.83, compared to $26.79 for AI. Consequently, the ROAS for human ads was 2.5x, a full 0.4 points higher than AI’s 2.1x. It seems the nuance in human-written copy resonated more deeply, even with identical targeting.

Paid Search Performance (Google Ads)

Total Search Ad Spend: $100,000 ($50,000 per track)

Metric AI Track Human Track
Impressions 8,200,000 7,900,000
Clicks 106,600 118,500
Click-Through Rate (CTR) 1.30% 1.50%
Conversions (Purchases) 1,600 2,100
Cost Per Conversion (CPC) $31.25 $23.81
Return on Ad Spend (ROAS) 1.8x 2.3x

The pattern observed in paid social largely held true for Google Search Ads. Human-written ad copy and headlines resulted in a CTR of 1.50%, surpassing AI’s 1.30%. The CPC for human ads was $23.81, significantly better than the $31.25 for AI ads. This led to a 2.3x ROAS for human ads versus 1.8x for AI. The conciseness and persuasive power of human-crafted search ad copy, especially for high-intent keywords, proved superior.

What Worked and What Didn’t

AI Track:

What Worked: The sheer speed and cost-efficiency of generating a large volume of diverse creative variants were undeniable. We could test dozens of headlines and image combinations in hours, something that would have taken days for the human team. This allowed for rapid iteration and identification of underperforming assets. The AI excelled at generating generic, visually appealing lifestyle shots that met basic brand aesthetic requirements. What Didn’t Work: AI struggled with capturing subtle brand nuances and emotional resonance. Many AI-generated images, while technically proficient, felt somewhat sterile or generic. The copy, despite detailed prompts, sometimes lacked the persuasive flair and unique brand voice that distinguishes a premium DTC brand. For instance, an AI-generated headline like “Sustainable Clothes for Modern Life” performed worse than a human-written “Crafted for Conscience: Improve Your Everyday with Thoughtful Apparel.” The lack of authentic storytelling in AI copy was a consistent weakness.

Human Track:

What Worked: The human team produced creatives with a stronger emotional connection and a clearer articulation of the brand’s unique selling propositions. Original photography featuring local models (from the Atlanta area, for example) and authentic settings resonated more strongly with the target audience. The copy was more engaging, often incorporating subtle calls to action and storytelling elements that AI couldn’t replicate. Their ability to adapt quickly to minor campaign shifts with fresh, relevant creative was also a significant advantage. What Didn’t Work: The primary drawback was the higher cost and slower production cycle. Each creative iteration required manual effort, limiting the volume of A/B tests we could run. This meant that while individual human creatives performed better, the sheer number of AI variants allowed for more aggressive testing to find “winners” in specific niches.

Optimization Steps Taken

Recognizing the strengths and weaknesses of each approach, we implemented a hybrid optimization strategy in the latter half of the campaign (weeks 5-8).

  1. AI for Ideation, Human for Refinement: We began using AI to generate 20-30 initial ad copy variations and image concepts. The human copywriter and designer then selected the most promising 5-7, refining them for tone, brand voice, and visual impact. This drastically cut down the human team’s ideation time, allowing them to focus on polishing.
  2. Performance-Based Allocation: We gradually shifted more budget towards the top-performing human-crafted ad sets, especially on Meta where visual storytelling is paramount. For Google Search, we used AI to generate long-tail keyword ad variations, which the human team then reviewed and optimized for relevance and Quality Score.
  3. Dynamic Creative Optimization (DCO): On Meta, we leveraged DCO, feeding both AI and human-generated elements into the system. This allowed the platform to dynamically assemble the best-performing combinations of headlines, body copy, and visuals based on real-time user engagement. This approach showed a 1.2 percentage point increase in CTR for the combined DCO sets compared to either pure AI or pure human-generated static ads.

The Verdict: A Hybrid Future

This campaign demonstrated that while AI offers undeniable advantages in speed and cost reduction for creative production, it currently struggles to match the nuanced persuasive power and emotional resonance of human-crafted ads. The average cost per conversion across both platforms was $29.02 for AI versus $22.32 for human ads, a significant difference. However, dismissing AI entirely would be short-sighted. The most effective approach emerged as a hybrid model. Using AI for rapid ideation, generating vast quantities of initial concepts, and handling repetitive tasks frees up human creatives to focus on strategic thinking, brand storytelling, and refining the most promising outputs. This allows for the best of both worlds: the efficiency and scalability of AI combined with the creativity and emotional intelligence of human talent. Marketers should view AI as a powerful assistant, not a complete replacement. The future of effective ad creative lies in intelligent collaboration between machine and mind.

Can AI fully replace human ad copywriters and designers?

Based on our campaign data, AI cannot fully replace human ad copywriters and designers, especially for brands requiring nuanced storytelling and strong emotional connection. While AI excels at generating high volumes of content quickly and cost-effectively, human creatives consistently produce ads with higher engagement rates and better conversion performance due to their ability to understand and convey complex brand values and emotional appeals.

What are the primary cost benefits of using AI for ad creative?

The primary cost benefits of using AI for ad creative stem from significantly reduced production time and labor costs. Our campaign saw a 40% reduction in creative production spend for the AI track, mainly by minimizing the need for extensive human design and copywriting hours, enabling rapid iteration and testing of numerous creative variants at a lower expense.

How can marketers best integrate AI into their ad creative workflow?

Marketers can best integrate AI by adopting a hybrid approach. This involves using AI tools for initial ideation, generating diverse concepts, and handling repetitive tasks, then having human creatives refine and optimize the most promising outputs. This strategy leverages AI’s efficiency for quantity and speed, while human oversight ensures brand consistency, emotional resonance, and strategic alignment.

Which performance metrics showed the biggest difference between AI and human ads?

The biggest differences were observed in Click-Through Rate (CTR), Cost Per Conversion (CPC), and Return on Ad Spend (ROAS). Human-crafted ads consistently achieved higher CTRs (e.g., 1.80% vs. 1.50% on social), leading to lower CPCs (e.g., $20.83 vs. $26.79 on social) and in the end a higher ROAS (e.g., 2.5x vs. 2.1x on social), demonstrating superior conversion efficiency.

What role does prompt engineering play in AI ad creative success?

Prompt engineering plays a critical role in AI ad creative success. The quality and specificity of the prompts directly influence the relevance and effectiveness of the AI-generated output. Detailed prompts that include brand guidelines, aesthetic preferences, target audience insights, and emotional tones are essential for guiding AI to produce creatives that are closer to human-level quality and align with campaign objectives.

Share
Was this article helpful?

David Jackson

Digital Marketing Strategist

David Jackson is a leading Digital Marketing Strategist with over 14 years of experience revolutionizing online presence for global brands. As the former Head of Performance Marketing at Zenith Digital Solutions and a Senior Strategist at Impact Media Group, David specializes in advanced SEO and content strategy, driving organic growth and measurable ROI. Her innovative methodologies have consistently placed clients at the forefront of their industries. She is the author of the influential white paper, 'The Algorithmic Shift: Adapting Content for Tomorrow's Search Engines'