Adapting Paid Ad Strategies to AI-Driven Search and Video
Marketing Against the Graingo watch the original →
the gist
AI Overviews have structurally inflated Google Search ad costs while reducing reach, creating a 12-18 month arbitrage window in YouTube Shorts and requiring a shift toward hyper-targeted LinkedIn campaigns.
The Impact of AI on Search Economics
AI Overviews have fundamentally disrupted paid search, with coverage increasing from 6% to over 50% of queries in 18 months. This shift has caused Google cost-per-click (CPC) rates to rise by 14% while paid impressions have dropped by 6%. Marketers should not abandon search, but they must reallocate budgets to account for the diminished reach and higher costs of traditional search placements.
Arbitrage Opportunities in YouTube Shorts
YouTube Shorts currently offers a significant arbitrage opportunity with a cost per thousand impressions (CPM) of $4.85. The market suffers from a supply-demand imbalance: while AI tools have made it easier for creators to produce high-quality video, many businesses have not yet built the internal workflows to match that output. This window is expected to close within 12 to 18 months as advertiser adoption catches up. To exploit this, marketers should adopt a tooling stack including HeyGen, Captions.ai, and Canva to rapidly iterate on video ad creative.
The LinkedIn Paradox and Creative Equations
LinkedIn remains a highly effective B2B channel, though it is often mismanaged due to a focus on high CPMs (approximately $60). The discrepancy in performance between marketers who find LinkedIn expensive and those who find it effective stems from targeting strategy. Effective campaigns avoid generic messaging to broad audiences, instead using specific messages for highly tailored segments. Success requires moving beyond CPM-based metrics to focus on creative equations that prioritize cost-per-click and customer acquisition efficiency.
The Loop Framework
To optimize ad performance, apply the four-stage Loop framework:
- Express: Generate initial concepts by feeding customer interviews and product data into an LLM to produce 5 to 10 script variations.
- Tailor: Narrow targeting to specific audience subsets that align with the chosen script.
- Amplify: Produce and publish the video ad.
- Evolve: Feed performance data back into the LLM to generate insights for the next iteration of the loop.