The Art of Manufacturing Irrelevance
Platforms don’t just recommend content—they manufacture disinterest. Learn about audience poisoning, CTR manipulation, and the Google Search Console mirage. Western Civilization Can't Be Honest #15
We are told the algorithm is a mirror. It reflects our desires, curates our tastes, and serves us what we “want” to see. This is the foundational myth of the engagement economy. It is a comforting lie. The reality is far more insidious: the algorithm is a gatekeeper with a financial incentive to manufacture disinterest, and it possesses the ultimate tool to do so—the ability to make content invisible without ever marking it as removed.
The discourse around algorithmic transparency is mired in a fatal flaw. We argue about “bias” and “moderation” as if the algorithm were a judge rendering a verdict on content. We ask, “Did this video violate the rules?” when we should be asking, “Did this video ever get a fair trial?” The platform does not need to ban a video to kill it. It merely needs to starve it of oxygen.
Consider the enigma of the Click-Through Rate (CTR). To the uninitiated, CTR is a simple metric: the percentage of people who see a thumbnail and click it. But in the hands of a recommendation engine, CTR is a weapon of obfuscation. The platform knows, with chilling accuracy, exactly who will click on a piece of content. It possesses terabytes of data on your viewing habits, your watch history, your pause points, and your scroll velocity. It knows, often better than you do, the exact demographic and psychographic profile that would find a specific video irresistible.
Yet, the platform deliberately chooses not to show that content to those people.
Allegedly
Instead, it engages in a practice we might call “Audience Poisoning.” It takes a niche but highly relevant video and briefly serves it to an audience it knows will reject it—users who have never watched a video on that topic, whose viewing histories are diametrically opposed to the subject matter. When these users inevitably scroll past, the CTR plummets. The algorithm then turns to the creator and declares, “See? No one wants this. Your content is bad. Your thumbnail is ineffective.”
This is not a failure of the algorithm; this is the algorithm functioning exactly as designed. It is a supply-chain management system. It is suppressing the supply of content that does not fit the platform’s preferred narrative for a given user, and it is using the manufactured rejection of that content as its justification.
In 2021, internal documents disclosed by Facebook whistleblower Frances Haugen showed that Meta deliberately “dialed down” the reach of civic and political content to reduce user complaints, admitting it could selectively reduce visibility without removing content entirely (The Wall Street Journal, The Facebook Files).
The “For You” page is the primary vector for this manipulation. It is sold as a personalized feed, a digital concierge predicting your tastes. In truth, it is a demand-management tool. The platform has finite “inventory” of user attention. It must allocate that attention to content that maximizes its primary metric: time on site. However, “time on site” is not merely about keeping you on the platform for thirty minutes; it is about keeping you on the platform and exposing you to the highest-margin advertisements.
A video about a complex geopolitical conflict might be deeply interesting to you. You are a history buff; you watch long-form documentaries daily. The platform knows this. But that video does not have a high-value ad inventory attached to it, or it features a controversial figure that makes brand-safety advertisers nervous. To solve this, the platform does not ban the video. It simply “disinterests” you in it. It hides the video behind a wall of clickbait about a celebrity breakup, knowing that if it gives you the option between the two in a scroll, you will pause on the celebrity drama, even if just for a second.
Multiple legal complaints have argued that platforms use such preference-shaping not merely to serve users but to enforce advertiser-friendly content norms. In a 2017 federal lawsuit, PragerU alleged that YouTube’s Restricted Mode—an algorithmic filter ostensibly designed to hide “potentially mature” content—was applied in a discriminatory manner to suppress its educational videos from certain viewers, effectively manufacturing disinterest at scale (PragerU v. Google). Similarly, a 2019 lawsuit filed by a group of LGBTQ+ creators claimed YouTube’s monetization algorithm unfairly demonetized and restricted their videos under the guise of advertiser-suitability, suppressing their reach without ever issuing a policy violation (Divino Group LLC v. Google LLC).
When the algorithm finally decides to surface the political video, it does so at 2:00 AM, when your engagement is low. It places it third in your feed, after two videos you are guaranteed to click on, so that the impression is counted but the click is not. The result is a suppressed CTR, a low average view duration (since you are groggy), and a statistical death sentence for the creator.
The crucial element of this system is its impenetrable opacity. When pressed on these issues, platforms point to their “transparency reports,” which detail how many videos were removed for policy violations. This is a masterful deflection. They are showing you the corpses they buried in the cemetery, while completely hiding the thousands of videos they poisoned in the hospital.
You cannot audit a ghost. You cannot scrutinize a non-event.
When a creator complains that their views have tanked, the platform responds with vague platitudes: “The algorithm changed,” or “Engagement is down.” There is no data provided on who the video was shown to. There is no data on why that specific subset of users was chosen over the creator’s core fanbase. The platform holds the monopoly on the data that proves its own malpractice.
Furthermore, the platforms are wise to the criticism. They utilize “A/B testing” to prove they are right. They will run an experiment where they show a “controversial” video to 1% of a target audience and a “safe” video to the other 99%. When the controversial video performs worse (because it is only shown to the 1% that are fatigued), the platform claims its algorithm is optimized for user satisfaction. This is circular logic. They define satisfaction by their own manipulated engagement metrics, and they control the experiment. They are the referee, the player, and the scorekeeper.
In a genuinely transparent ecosystem, a creator would be able to see their “Audience Match” score. They could see a heatmap of the world, showing exactly where their videos are being suppressed and where they are being surfaced. They would be able to see the demographics of the non-clickers and compare them to the demographics of the engaged viewers.
But we do not have that, and we will never have that, because that transparency would ruin the business model. The platforms are not in the business of connecting people with content; they are in the business of controlling the pipe through which content flows.
If Search Was A Feature And Not A Bug
What happens on video platforms is not unique; it is merely the most visible expression of a logic that governs the entire infrastructure of attention. Google Search Console operates on an identical principle of manufactured statistical fatalism, dressed in the clinical language of “impressions” and “average position.”
When a web page appears in a Google search result, that counts as an impression. If no one clicks it, the click-through rate drops. Google’s documentation frames a low CTR as a signal that the page is not satisfying user intent—a gentle, data-driven nudge that perhaps the content needs improvement. But this narrative omits a crucial, deliberate ambiguity: Google decides which queries trigger the impression. The platform can show your meticulously researched article on a niche historical event for a vague, high-volume keyword tangentially related to the topic—knowing full well that the searcher’s intent is commercial, navigational, or utterly mismatched. When the inevitable scroll-past occurs, Google records another “failure” for your content, feeds it into the ranking algorithm, and demotes your page for that query. The platform has not removed your page from the index; it has suffocated its visibility by forcing it to compete with its own manufactured audience of misaligned intent.
Even the impression data itself is unreliable. Google Search Console aggregates impressions from search result pages where your link may be buried below the fold or displayed only after a series of competitor ads, yet the impression is logged exactly the same as if it sat at the top of the screen. You, the creator, are shown an aggregate chart—a clean, upward-trending line of “impressions” and a downward-sloping CTR percentage—and told that the numbers speak for themselves. But you cannot see the query-level context, the true position on the page, the modality (image, video, web), or the user’s prior search history that shaped the result set. Just as a video platform hides the demographics of the non-clickers, Google Search Console hides the intent of the searcher. The opaque impression becomes a weaponized metric: “Your page got ten thousand impressions this month, yet only a 0.5% CTR. It must be irrelevant.” This is not measurement; it is entrapment by data.
The parallel is exact. Where social platforms use Audience Poisoning to show the wrong video to the wrong user and kill a creator’s reach, Google uses Intent Mismatching to show the right page to the wrong query and kill a publisher’s rankings. In both systems, the platforms retain absolute discretion over the initial distribution while presenting the resulting engagement metrics as organic verdicts. The creator sees only the fatal statistic, never the experimental condition that produced it. And in both cases, the opacity is structural: revealing the mismatch would expose the platform’s role in manufacturing disinterest, turning a gatekeeper’s discretionary power into a measurable and contestable act.
The myth of the algorithm is that it is a dispassionate arbiter of taste. The reality is that it is a partisan gatekeeper, leveraging its monopoly on data to manufacture consent and, more importantly, manufacture disinterest. When we accept CTR and “For You” metrics as gospel, we are not engaging in a critique of the system; we are acting as unpaid data-validation clerks for a machine that has already decided who is worthy of being seen.