Marketing Attribution & AI Analytics with Keith Perhac

The podcast discusses the evolving landscape of marketing attribution in the era of AI and increasing bot traffic, where over half of internet traffic now comes from non-human sources and most Google searches result in no external clicks. Traditional attribution models are becoming less effective as AI tools like ChatGPT aggregate information without directing users to websites, making it difficult to track how consumers discover brands. The conversation emphasizes the need for marketers to adapt by focusing on both first-touch and last-touch attribution, leveraging more controllable channels like email and websites, and combining objective tracking data with subjective customer feedback to gain accurate insights.

A key theme is the importance of human-centric marketing strategies despite the rise of AI. The discussion highlights that while AI can process vast amounts of data, it lacks the ability to interpret context or drive meaningful action without human guidance. Marketers are encouraged to prioritize first-party data collection, use AI to repurpose human-created content rather than generate it from scratch, and focus on segmentation - especially in B2B contexts where audience sizes are smaller. Additionally, the podcast explores how AI can streamline funnel optimization by identifying high-impact audience segments and personalizing content based on behavioral traits, while cautioning against overcomplication and stressing the continued value of expert interpretation in making data actionable.

• What if you started tracking first-touch attribution from AI-driven discovery for your own product?

  • Move: Add a simple onboarding question (e.g., "How did you first hear about us?") with options including "AI assistant (e.g., ChatGPT, Claude)" and log responses in your CRM or spreadsheet.
  • Why Now?: With over 50% of traffic being bots and AI agents, and zero-click searches rising, traditional tracking misses early discovery - especially when AI surfaces your content without a direct click.
  • Expected Upside: Identify untapped top-of-funnel channels; refine SEO/content strategy to optimize for AI visibility and capture users at first contact.

• What if you built a single high-converting funnel tailored to your best-performing audience segment?

  • Move: Analyze existing conversion data by segment (e.g., job title, industry, behavior), pick the one with the highest conversion rate, and rebuild your core funnel (lead magnet, email sequence, offer) specifically for them.
  • Why Now?: AI can quickly analyze behavioral outliers, but acting on one high-impact segment first prevents overwhelm and delivers faster ROI than broad, generic funnels.
  • Expected Upside: Increase conversion rates by 20 - 50% in that segment; create a profitable baseline to expand from, reducing customer acquisition cost.

• What if you repurposed one piece of human-created video content into 10+ platform-specific assets using AI?

  • Move: Record a 10-minute video answering a core customer question, then use AI tools to generate short clips, LinkedIn posts, blog summaries, email snippets, and social captions across 5+ platforms.
  • Why Now?: Video is the most repurposable format, and AI drastically cuts distribution time - critical as attention fragments across platforms and AI-generated content floods the web.
  • Expected Upside: Maintain consistent, authentic presence across channels with 80% less effort; improve discoverability where your audience spends time (e.g., LinkedIn, YouTube, email).

• Implement first-touch and last-touch attribution tracking in your analytics setup to identify both initial discovery sources and final conversion channels.
• Add simple onboarding questions (e.g., "Where did you hear about us?") to collect first-party data directly from customers and validate tracking insights.
• Focus on controllable marketing channels like email and owned content; build segmented email workflows tailored to audience types (e.g., developers vs. marketers) using behavioral and self-reported data.
• Create one core piece of human-generated content per week (e.g., a video or podcast) and use AI tools to repurpose it into blog posts, social media snippets, and email content across platforms.
• Use AI to analyze conversion funnels by segment (e.g., industry, persona), identify the lowest-performing funnel, and prioritize optimizing or sunsetting it before scaling to others.