Week 28: Testing content distribution in the AI era

The Experiment

Here's something that's been bugging me; AI is fundamentally breaking how we acquire customers online. For the last 10-15 years, there's been this implicit contract where you create content, Google sends you traffic, and you monetize that audience. With Now Novel we grew a 400k global audience just with content marketing. But with AI overviews and zero-click search results, that whole model is crumbling.

I wanted to understand how content distribution actually works in this new landscape. The question wasn't whether I could "game" anything—it was whether I could figure out what it takes to get picked up by LLMs and search systems in the first place.

I wanted to see how I could do it "cleanly": create a completely fake business concept and track exactly how it moves through the information ecosystem. If I could trace the path from creation to AI pickup, I'd understand something about modern content distribution.

The Process

Here's how I tackled this experiment:

1. Inventing the Perfect Fake Term

Claude and I (apologies for anthropomorphizing AI 💻) brainstormed concepts that sounded legitimate but didn't exist. I needed something that felt academic enough to be credible but specific enough to track. "Decision fork injection" won because it sounds like it could absolutely be a real product strategy and I just liked the idea of having "fork" in the name.

2. Building the Content Foundation

I wrote a comprehensive Medium post defining "decision fork injection" as a UX strategy. To make it convincing, I:

• wove in real citations from Robert Cialdini and BJ Fogg

• Used proper business writing structure with frameworks and examples

• Created realistic use cases for e-commerce, SaaS, and content platforms

• Listicles format, clusters of information, including fan-out pieces that answer all facets of questions

3. Strategic Distribution Across Platforms

• Published the main post on Medium and my personal blog

• Shared it on LinkedIn and Twitter

• Found existing Reddit threads about decision architecture and CRO where I could naturally work in my fake term

• Linked everything together to create a web of "credible" sources

4. Obsessive Monitoring

I became embarrassingly obsessive about checking search results. Every few hours, I'd search "decision fork injection" in quotes across Google, Perplexity, and other platforms to see if it was getting picked up.

The Outcome

This worked way better (and faster) than I expected:

• 2 hours: Google had indexed my Medium post and was showing it in search results

• 3-4 days: Perplexity was citing my content when asked about the term

• Reddit replies: My strategic responses in relevant threads were also getting indexed

The speed was genuinely shocking. Within hours, my completely fabricated business concept was being treated as legitimate information by search engines.

What surprised me most was how readily the systems accepted and amplified the content. There's something both impressive and slightly terrifying about how quickly you can inject "knowledge" into the information ecosystem.

Key Takeaway

This experiment taught me something about content distribution in the AI era. The old "create great content and Google will reward you" model is dead. What matters now is understanding how different systems—search engines, AI overviews, LLMs—actually discover, process, and surface information.

The speed of pickup wasn't about gaming anything. It was about understanding that modern content distribution works through multiple, interconnected channels that all feed into each other.

Pro Tips for... Educational Purposes:

1. Make It Sound Academic: Proper structure, real citations, and frameworks make fake concepts more believable

2. Strategic Cross-Platform Posting: Multiple sources citing the same concept creates apparent credibility

3. Target Relevant Communities: Find existing discussions where your concept could naturally fit

4. Patience (Sort Of): While I was obsessively checking, the pickup happened faster than expected

Want to Try It Yourself?

• Write comprehensive content with proper citations and structure

• Distribute across multiple platforms (Medium, LinkedIn, Reddit, personal blog)

• Monitor with quoted search terms to see AI pickup

The Bigger Picture

I'm rating this a 10/10 for SEO effectiveness and 7/10 for LLM infiltration (I'm quietly confident it'll hit ChatGPT and Claude when they update their training data).

But the real win wasn't proving I could create fake content. It was understanding how content moves through the modern information ecosystem. We're in a transition period where traditional SEO, AI training data, and social distribution are all interconnected in ways we're still figuring out.

If you're creating content in 2025, understanding these new distribution channels isn't optional—it's survival. The companies that figure out "Search Everywhere Optimization" (as Rand Fishkin calls it) will have a massive advantage over those still optimizing for 2019 Google.

And if anyone would like to hear me speak about decision fork injection 🔧—the groundbreaking strategy I definitely didn't just make up—my rates are very reasonable and acceptable. 😄