Beyond the Click: How AI is Rewriting the Rules of Revenue Optimization

Conversion Rate Optimization (CRO) is often misunderstood as a design discipline, but Ned MacPherson argues it is a data-driven revenue engine. By shifting focus from aesthetics to post-click behavior, brands can unlock hidden value. This conversation reveals that the most effective optimizations, such as AI-driven pricing experiments and LLM-friendly data structures, often live in the parts of a website that humans ignore but algorithms prioritize. For marketers and business owners, the advantage lies in moving away from copying competitor designs toward building systems that respond to real-time market demand. This approach requires the patience to test foundational mechanics, but it creates a competitive moat that design-first competitors cannot easily replicate.

The Hidden Cost of Competitor-Led Design

The prevailing industry approach to CRO is flawed: brands copy the design elements of successful competitors, assuming those choices are the cause of their success. MacPherson identifies this as a dangerous shortcut.

Most folks in this industry will look at a site and they will look at the structure or the homepage and the navigation and the PDPs and the cart and what have you and then they will look at competitors and they are kind of just making the blind assumption that that site does really well analytically.

-- Ned MacPherson

This design-centric approach creates a feedback loop of false confidence. When a brand copies a competitor and sees a conversion win, they attribute it to the design change. In reality, they are often shooting with a blindfold on. The systemic risk is that teams embed themselves in a process of imitation rather than empirical diagnosis, leading to wasted effort on high-visibility pages that may already be performing at their peak.

Where Immediate Pain Creates Lasting Moats

Systems thinking requires looking beyond the add to cart event. MacPherson highlights that for high-consideration purchases, like luxury jewelry, the primary friction point is often the user desire to validate price against the broader market.

Instead of forcing users to leave the site to compare prices, MacPherson built an AI agent that scours the web every two hours to display where the brand price sits relative to the market mean. This addresses the pricing objection immediately. The downstream effect is a shift in user focus: once the pricing anxiety is removed, the user is free to engage with the brand quality and curation value propositions. This creates a competitive advantage because it addresses the user hidden intent, which is to compare, rather than trying to suppress it.

Optimizing for the Invisible User: The LLM

A non-obvious consequence of the rise of Generative AI is the need for Generative Engine Optimization (GEO). MacPherson notes that while humans have an aversion to scrolling and dense data, LLMs thrive on it.

We have actually built, I will admit, not very attractive but very LLM friendly tables, us versus them comparison charts, data charts, things that things like LLM bots love, right? Whereas the human might look at it and be like this is overwhelming, you are throwing too much at me, not for LLM.

-- Ned MacPherson

By placing structured, data-heavy tables at the bottom of a page, where human engagement is naturally low, brands can satisfy the requirements of LLMs for search and recommendation without cluttering the human user experience. This is a systems-level play: you are bifurcating your site architecture to serve two different masters simultaneously.

The Autonomous Revenue Loop

The most sophisticated application of this thinking is the use of AI agents to manage SKU-level pricing. In environments with thousands of SKUs, human teams typically focus only on the top performers. MacPherson uses AI to monitor sales velocity in real-time against predictive thresholds. If a product lags, the AI autonomously injects a discount experiment. If it sells too fast, it can test price increases. This creates a self-optimizing system that captures revenue opportunities in the iceberg of products that human teams are too busy to manage.

Key Action Items

• Audit Your Forms (Immediate): If you are in a service-based business, stop asking for excessive information. Remove non-essential fields like phone numbers or how did you hear about us to reduce friction. You can gather those details during the follow-up call.
• Implement Post-Click Automations (Next Quarter): If your conversion process involves a lead capture, ensure your follow-up automation is as robust as your landing page. A high lead count is irrelevant if the system fails to bridge the gap to the actual appointment.
• Identify Your Invisible Data (Next 6 Months): Identify the data points that prove your value, such as comparison charts, specs, or quality metrics. Move these into structured tables at the bottom of your product pages to improve your visibility for LLMs.
• Shift to Empirical Testing (Ongoing): Stop redesigning based on pretty competitor sites. Use heatmaps and funnel analysis to identify where your site is actually weak. Create alternative hypotheses to test against your current performance.
• Explore Autonomous Pricing (12-18 Months): If you manage a large SKU count, investigate AI-driven agents that can autonomously test price elasticity via coupon injection. This requires significant engineering setup but creates a massive advantage in operational efficiency.