How to Build Your Own Predictive SEO Pipeline

<p>Has anyone else seen the studies about AI reducing critical thinking that have been making rounds on LinkedIn?</p>
<p><strong>Prepared to be disappointed.</strong></p>
<p>What they researched ≠ What they claimed.</p>
<p>There&#39;s a lot of irony there... If you apply critical thinking to the study itself, you find this out yourself.</p>
<hr>
<h2>🚩 The Methodological Problems</h2>
<p>Here are the key issues with this viral research, some of which are acknowledged in the study&#39;s limitations section, others not:</p>
<h3>Issue #1: Self-Reported Survey Data</h3>
<p>The study measures &quot;critical thinking&quot; using <strong>self-reported survey questions</strong>. No tests. No experiments. Just how people <em>think</em> about their own critical thinking skills.</p>
<p>This is a fundamental flaw. Self-perception of cognitive ability is notoriously unreliable. The <a href="https://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect">Dunning-Kruger effect</a> exists precisely because people are bad at accurately assessing their own cognitive abilities.</p>
<p>If you want to measure whether AI affects critical thinking, you need to actually <em>test</em> critical thinking — not ask people how they feel about it.</p>
<hr>
<h3>Issue #2: The Questions Measure Trust, Not Thinking</h3>
<p>The actual &quot;critical thinking questions&quot; in the survey are more about <strong>how much you trust AI</strong> and less about measuring your ability to think critically.</p>
<p>Some examples from their methodology:</p>
<blockquote>
<p>&quot;How often do you critically evaluate sources?&quot;</p>
</blockquote>
<blockquote>
<p>&quot;How confident are you in spotting fake news?&quot;</p>
</blockquote>
<blockquote>
<p>&quot;Do you question the motives behind AI recommendations?&quot;</p>
</blockquote>
<p>These questions measure <strong>attitudes toward verification</strong> and <strong>trust in AI systems</strong> — not actual critical thinking ability.</p>
<p>There&#39;s a massive difference between asking &quot;Do you think you&#39;re a critical thinker?&quot; and testing whether someone can identify logical fallacies, evaluate evidence, or construct sound arguments.</p>
<hr>
<h3>Issue #3: The Foundation Paper Problem</h3>
<p>The researchers claim their methodology is based on another research paper about measuring critical thinking. But there are two problems:</p>
<h4>🚩 Issue #3.1: Wrong Audience</h4>
<p>That foundational paper was <strong>designed for children</strong>, while this study was applied to adults. Cognitive assessment tools don&#39;t just transfer across age groups without validation.</p>
<h4>🚩 Issue #3.2: They Don&#39;t Follow the Methodology</h4>
<p>They don&#39;t actually follow the methodology of the originating paper. They adapted it (loosely), which undermines the claim that they&#39;re building on validated research.</p>
<hr>
<h2>🤔 What I Think They Actually Found</h2>
<p>Despite the sensationalist framing, there <em>are</em> some legitimate insights buried in this research:</p>
<h3>1. People Are Trusting AI Outputs Without Verification</h3>
<p>More people are trusting AI outputs instead of doing their own research or verification — and <strong>that&#39;s the part worth paying attention to</strong>.</p>
<p>Not because AI is inherently harmful… but because it shows our deep reliance on these systems for information.</p>
<p>I don&#39;t fact-check my calculator when I plug in 5 + 5. And increasingly, people aren&#39;t going to fact-check their LLM of choice for gathering information either.</p>
<p>That&#39;s a behavioral shift worth understanding — but it&#39;s not the same as &quot;AI is eroding critical thinking.&quot;</p>
<hr>
<h3>2. There&#39;s Something Here Worth Researching More</h3>
<p>Despite the sensationalist headline, there were some interesting findings worth exploring in more detail with better methodology.</p>
<p>The question of how AI tools affect our cognitive patterns is genuinely important. We just need better research to answer it.</p>
<hr>
<h2>📊 The Broader Point</h2>
<p>Before we share misinformation with headlines saying <strong>&quot;AI is eroding critical thinking,&quot;</strong> let&#39;s model the skill we&#39;re worried about losing.</p>
<p>This is exactly the kind of content that goes viral because it confirms existing fears. The headline is scary. The conclusion feels intuitive. And most people won&#39;t dig into the methodology.</p>
<p>But if we&#39;re genuinely concerned about critical thinking in the AI age, the best thing we can do is demonstrate it ourselves:</p>
<ul>
<li><strong>Question sensationalist claims</strong> — even (especially) when they align with our priors</li>
<li><strong>Read the methodology</strong>, not just the abstract</li>
<li><strong>Distinguish between what was measured and what was claimed</strong></li>
<li><strong>Consider alternative explanations</strong> for the findings</li>
</ul>
<hr>
<h2>📎 Read the Original Study</h2>
<p>If you want to evaluate this yourself (and I encourage you to), here&#39;s the original paper:</p>
<p><strong><a href="https://www.mdpi.com/2075-4698/15/1/6">AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking</a></strong></p>
<p>Apply some critical thinking to it. See what you find.</p>
<hr>
<h2>💭 Final Thoughts</h2>
<p>AI tools are powerful. They&#39;re changing how we work, research, and make decisions. There are legitimate questions about how reliance on these systems affects our cognitive habits.</p>
<p>But answering those questions requires rigorous research — not surveys about how people <em>feel</em> about their thinking abilities.</p>
<p>The irony of spreading a poorly-constructed study about critical thinking… while failing to apply critical thinking to the study itself… is almost too perfect.</p>
<p>Let&#39;s be better than that.</p>


That Viral AI Study About Critical Thinking? Apply Critical Thinking to It.

<p><strong>Meta survives another day.</strong></p>
<p>The U.S. District Court for the District of Columbia found in favor of Meta Platforms, Inc. (formerly Facebook), ruling that the Federal Trade Commission (FTC) failed to prove that Meta holds a monopoly in the relevant product market.</p>
<hr>
<h2>📋 The Case</h2>
<p>The FTC alleged that Meta monopolized the <strong>&quot;personal social networking&quot; (PSN)</strong> market by acquiring upstart rivals Instagram and WhatsApp. The argument: Facebook saw these companies as existential threats and bought them to eliminate competition rather than compete on merit.</p>
<p>It&#39;s the classic &quot;buy or bury&quot; antitrust theory — and on paper, it looked like the FTC had a strong case. The internal communications from Facebook executives during the Instagram and WhatsApp acquisitions were... not subtle.</p>
<hr>
<h2>⚖️ The Ruling</h2>
<p>The Court, however, found that the social media landscape has <strong>evolved significantly</strong> since the suit was filed.</p>
<p>The key issue came down to <strong>market definition</strong> — arguably the most important (and most contested) element in any antitrust case.</p>
<p>The FTC defined the market narrowly: &quot;personal social networking services.&quot; In that market, Meta looked dominant.</p>
<p>But the Court determined that Meta competes in the <strong>broader social media market</strong>, which includes fierce rivals like:</p>
<ul>
<li><strong>TikTok</strong> — which has fundamentally reshaped how people consume and create content</li>
<li><strong>YouTube</strong> — still the dominant video platform with massive engagement</li>
<li><strong>Other platforms</strong> competing for the same user attention and advertiser dollars</li>
</ul>
<p>When you zoom out to that broader market, Meta&#39;s position looks a lot less like a monopoly and more like one major player among several.</p>
<hr>
<h2>🤔 What This Means</h2>
<h3>For Meta</h3>
<p>This is a major win. The company can continue operating Instagram and WhatsApp as integrated parts of its business without the threat of forced divestitures — at least from this case.</p>
<h3>For Antitrust Enforcement</h3>
<p>This ruling highlights how difficult it is to bring antitrust cases in tech markets that evolve rapidly. By the time a case gets to trial, the competitive landscape may look completely different than when the alleged anticompetitive conduct occurred.</p>
<h3>For the Industry</h3>
<p>The ruling reinforces that <strong>market definition is everything</strong> in antitrust. If regulators define markets too narrowly, they may overstate dominance. If courts define markets too broadly, genuinely anticompetitive behavior may go unchecked.</p>
<hr>
<h2>💭 The Bigger Question</h2>
<p>The interesting question isn&#39;t whether Meta is a monopoly <em>today</em> — it&#39;s whether the acquisitions of Instagram and WhatsApp <strong>prevented competition</strong> that would have otherwise emerged.</p>
<p>That counterfactual is incredibly hard to prove in court. What would Instagram have become without Facebook&#39;s resources? Would WhatsApp have remained independent? Would other competitors have filled the void?</p>
<p>We&#39;ll never know. And that uncertainty worked in Meta&#39;s favor.</p>
<hr>
<h2>📎 What&#39;s Next</h2>
<p>The FTC could appeal, but the ruling deals a significant blow to their case. The market definition issue will be hard to overcome on appeal.</p>
<p>For now, Meta lives to fight another day.</p>


Meta Survives Another Day: FTC Antitrust Case Dismissed

<p>After a year and a half of building what I genuinely believe was one of the most ambitious and transformative technologies in the SEO space, I made the difficult decision to shut down <strong>Kixely</strong>.</p>
<p>The mission was bold: <strong>reverse-engineer Google’s algorithm</strong> well enough to predict how changes to a website would impact rankings — finally giving companies a way to tie SEO actions directly to revenue, with real, data-driven precision.</p>
<p>The problem was one the industry had been stuck with for decades:<br>no one truly understood how Google’s own machine learning systems would respond to content, technical, or structural changes.  </p>
<p>So we built a platform capable of modeling Google’s ranking behavior across millions of pages and more than <strong>200 different signals</strong> — from content quality, to topical relevance, to backlink profiles, to speed, layout, and internal linking. And for a while, it worked better than even we expected… reaching <strong>up to 75% predictive accuracy</strong>.</p>
<hr>
<h2>⚙️ What Kixely Actually Did</h2>
<p>Kixely analyzed millions of pages across a brand and its full competitive set, then trained a machine learning model to mimic Google’s ranking patterns across hundreds of signals, including:</p>
<ul>
<li><strong>Content quality</strong> (structure, depth, coverage, semantic breadth)  </li>
<li><strong>Backlink characteristics</strong> (authority, diversity, recency, relevance)  </li>
<li><strong>Topical information gain</strong> at the page and domain level  </li>
<li><strong>Page-level signals</strong> (speed, layout, internal linking, UX patterns)</li>
</ul>
<p>With a fully trained model, we could run <strong>instant, at-scale SEO simulations</strong>:</p>
<ul>
<li><em>What if we publish a new product page?</em>  </li>
<li><em>What if we improve internal linking across our top URLs?</em>  </li>
<li><em>What if we improve speed by 20%?</em>  </li>
<li><em>What if we rewrite or expand a key piece of content?</em></li>
</ul>
<p>And in seconds, the system predicted not just ranking shifts — but the <strong>business value</strong> of those changes.</p>
<p>We built entire infrastructures around this: large-scale crawlers capable of scanning tens of millions of pages per day, pipelines running on Azure, deep integrations with SerpApi and Moz, and provisional patents around our modeling and data-processing methodology.</p>
<p>Only one company in the world — MarketBrew — seemed to be attempting anything comparable.</p>
<p>And yet, despite the technology working, the business ultimately didn’t.</p>
<hr>
<h2>💡 What We Learned</h2>
<h3>1️⃣ Enterprise sales is brutally slow</h3>
<p>We priced between $20K–$25K/year — reasonable for the sophistication of the tech. But enterprise SEO teams are chronically under-resourced, even inside Fortune 500s.  </p>
<p>I ran all sales myself — from cold outreach to demos — and learned firsthand how political, multi-layered, and slow enterprise cycles are. Deals stalled over budget freezes, timing issues, or shifting priorities.  </p>
<p>Without someone dedicated to running enterprise sales full-time, we couldn&#39;t scale both the tech and the pipeline.</p>
<hr>
<h3>2️⃣ The TAM was too small</h3>
<p>Reverse-engineering Google requires <strong>massive datasets</strong>, which only large enterprises have. That inherently limited our addressable market.</p>
<p>We explored a mid-market version, but the compute requirements made it impossible without losing predictive accuracy.</p>
<p>The companies who loved us were the ones least able to operationalize us.</p>
<hr>
<h3>3️⃣ Customer interviews were difficult at this level</h3>
<p>Enterprise SEOs are extremely hard to get time with. Conferences like brightonSEO were the best conversations — deep, technical, high-signal discussions — but formal interviews or follow-ups were slow and complex.</p>
<p>Even when teams loved the tech, they’d say:  </p>
<blockquote>
<p>“We need buy-in from multiple teams before moving forward.&quot;</p>
</blockquote>
<p>That is not quick to solve.</p>
<hr>
<h3>4️⃣ We over-invested in technology too early</h3>
<p>Kixely wasn’t something we could fake with a scrappy MVP. It had to be predictive or there was no point.  </p>
<p>But that meant we spent our early months on model accuracy, infrastructure, and scale, instead of validating market fit early.</p>
<p>We built too much before proving enough.</p>
<hr>
<h3>5️⃣ Agencies were intrigued but ultimately not a great fit</h3>
<p>Their margins are tight, and most already sell “SEO as a service.” An expensive tool only made sense if it helped sell more or charge more — and that was never consistently provable.</p>
<p>The value was strong, but the segment wasn&#39;t.</p>
<hr>
<h3>6️⃣ Impact ≠ adoption</h3>
<p>The tech worked.<br>But teams struggled to integrate it into real workflows. Some wanted fewer data layers, not more. Others needed simpler, more actionable insights.</p>
<p>We built an incredibly advanced engine… when many teams wanted a clearer map.</p>
<p>If I could rewind, I’d flip that: lead with immediate utility, then layer in the tech.</p>
<hr>
<h2>🔁 What I’d Do Differently</h2>
<h3>1️⃣ Bring on a sales-driven cofounder</h3>
<p>Enterprise sales requires daily momentum, relationship-building, and political navigation. I needed someone dedicated to that motion.</p>
<h3>2️⃣ Lead with value, not innovation</h3>
<p>Accuracy mattered — but clarity mattered more. I’d invest earlier in simplifying insights and showing immediate wins.</p>
<h3>3️⃣ Pivot sooner</h3>
<p>We identified promising pivots from user feedback, but by the time those became obvious, we were already 18 months deep into the original direction.</p>
<hr>
<h2>🙏 Gratitude</h2>
<p>I had the opportunity to work with brilliant engineers and data scientists, and together we built something that — I believe — genuinely advanced the state of predictive SEO. Speaking at brightonSEO and getting real feedback from world-class SEOs was a career highlight.</p>
<hr>
<h2>🚀 What’s Next</h2>
<p>I’ve shifted to a simpler, more scalable problem with a much broader audience:</p>
<p><strong>InterviewDroid</strong> — a platform that automates client and expert interviews using AI calling agents, then turns those conversations into creative briefs, blog posts, and marketing collateral. - <a href="https://www.interviewdroid.com">InterviewDroid</a></p>
<p>One 15-minute call → a week’s worth of marketing content.</p>
<p>It’s accessible, useful to SMBs and agencies, and far easier to operationalize than predictive SEO.</p>
<p>I’m also considering <strong>open-sourcing parts of Kixely</strong> so others can learn from (or build upon) what we created.</p>
<p>If you&#39;re interested in the tech, the research, or the journey — reach out. I’d love to chat.</p>


Why I Shut Down Kixely — And What's Next

<p>🚨 <strong>If AI-written content is a core part of your content strategy...</strong></p>
<p>...then you may be making the biggest mistake of 2025 and beyond.</p>
<p>Cyrus Shepard recently shared a study by Graphite that found some eye-opening numbers:</p>
<ul>
<li><strong>Over 50%</strong> of blog content is now AI-generated</li>
<li>Only <strong>14%</strong> of AI content appears in search results</li>
<li>Sample size: 31,000+ pages analyzed</li>
</ul>
<p>If your company is still investing in human-written content, this may initially sound like bad news. Everyone else is using AI to pump out content at scale — how can you possibly compete?</p>
<p><strong>But it&#39;s actually very, very, very good for you.</strong></p>
<hr>
<h2>🎯 The Simple Reason Why</h2>
<p><strong>The bigger the problem → the more resources Google will spend to solve it.</strong></p>
<p>Think back to the history of SEO manipulation:</p>
<h3>Link Farms (Early 2000s)</h3>
<p>Remember when you could buy thousands of backlinks from link farms and watch your rankings soar? Google invested heavily in link quality analysis, eventually releasing the Penguin algorithm that devastated sites relying on artificial links.</p>
<h3>Keyword Stuffing (Mid 2000s)</h3>
<p>&quot;Seattle SEO Seattle SEO expert Seattle SEO company Seattle SEO services Seattle&quot; — this used to work. Until Google got smart enough to recognize (and penalize) unnatural keyword density.</p>
<h3>Private Blog Networks (2010s)</h3>
<p>Building networks of sites to pass link equity seemed like a hack. Until Google systematically identified and devalued these networks, sometimes penalizing entire portfolios overnight.</p>
<h3>Content Farms (Early 2010s)</h3>
<p>Sites like Demand Media and eHow pumped out massive volumes of low-quality, keyword-targeted content. The Panda algorithm update wiped many of them off the map.</p>
<hr>
<h2>🤖 AI Content Is No Different</h2>
<p>Every time people come up with ways to &quot;cheat&quot; the system, Google invests heavily to:</p>
<ol>
<li><strong>Detect</strong> the manipulation at scale</li>
<li><strong>Systematically remove</strong> that content from search results</li>
<li><strong>Penalize</strong> bad actors</li>
</ol>
<p>AI content is following the exact same pattern.</p>
<h3>What Google&#39;s Rating Guide Says</h3>
<p>Google&#39;s Search Quality Rating Guidelines have a very clear stance on AI content — and it&#39;s not good.</p>
<p>The guidelines emphasize:</p>
<ul>
<li><strong>E-E-A-T</strong>: Experience, Expertise, Authoritativeness, Trustworthiness</li>
<li><strong>First-hand experience</strong> as a ranking signal</li>
<li><strong>Original insights</strong> over regurgitated information</li>
<li><strong>Author credibility</strong> and demonstrable expertise</li>
</ul>
<p>AI-generated content, by definition, cannot have first-hand experience. It cannot have genuine expertise. It&#39;s synthesizing existing information — not creating new knowledge.</p>
<hr>
<h2>📈 The Human Content Advantage</h2>
<p>If you&#39;re still investing in human-written content, I believe you&#39;re sitting on a <strong>massive ranking advantage</strong> that hasn&#39;t yet fully materialized.</p>
<p>Here&#39;s why:</p>
<h3>1. Scarcity Creates Value</h3>
<p>As AI content floods the internet, genuinely original human insights become rarer — and more valuable to both readers and search engines.</p>
<h3>2. Google Is Getting Better at Detection</h3>
<p>The same AI technology that makes content generation easy also makes detection easier. Google has access to the same (and better) language models. They know what AI output looks like.</p>
<h3>3. The Quality Gap Will Widen</h3>
<p>As more competitors shift to AI-only strategies, the quality of their content will converge to &quot;average.&quot; Human-written content with real expertise will stand out more than ever.</p>
<h3>4. E-E-A-T Signals Matter More</h3>
<p>Author pages, credentials, demonstrable experience, original research — these signals that AI cannot replicate are becoming increasingly important ranking factors.</p>
<hr>
<h2>🔮 What This Means For Your Strategy</h2>
<h3>If You&#39;re Using AI Content:</h3>
<p>Don&#39;t panic — but don&#39;t double down either. Consider:</p>
<ul>
<li>Using AI for research and drafts, not final content</li>
<li>Having human experts review, enhance, and add original insights</li>
<li>Focusing AI efforts on internal content, not SEO content</li>
<li>Building genuine author expertise signals</li>
</ul>
<h3>If You&#39;re Still Writing Human Content:</h3>
<p><strong>Stay the course.</strong> You&#39;re building a moat that will become more valuable as Google&#39;s detection improves.</p>
<p>Focus on:</p>
<ul>
<li>Original research and data</li>
<li>First-hand experience and case studies  </li>
<li>Expert opinions with demonstrable credentials</li>
<li>Content that AI literally cannot create (original interviews, proprietary data, unique perspectives)</li>
</ul>
<hr>
<h2>💃🕺 Human Writers Rejoice</h2>
<p>The content arms race isn&#39;t about who can produce the most words the fastest.</p>
<p>It&#39;s about who can produce content with genuine value, real expertise, and original insights.</p>
<p>That&#39;s a race humans can still win.</p>
<p>The companies investing in human-written content today are building competitive advantages that will compound as Google continues to refine its ability to distinguish authentic expertise from synthetic content.</p>
<p>The 50% of content that&#39;s AI-generated? Most of it is competing for the same 14% of search visibility.</p>
<p>The other 86%? That&#39;s where the opportunity lies.</p>
<p><strong>Human writers, your moment is coming.</strong></p>


Seattle's Favorite Digital Marketing Blog

How to Build Your Own Predictive SEO Pipeline

More Articles

That Viral AI Study About Critical Thinking? Apply Critical Thinking to It.

Meta Survives Another Day: FTC Antitrust Case Dismissed

Why I Shut Down Kixely — And What's Next

Why AI-Written Content May Be The Biggest SEO Mistake of 2025

Podcast Guest: Predicting SEO Success Using AI on The Simple and Smart SEO Show