AI Killed The SEO Hack
Gannett's decision to kill Reviewed is an example of Google's continued bulldozing of certain coverage — but it comes at a moment when trusted publications have more leverage than ever
Anxiety is a funny conundrum. It is the fear of what isn’t. It’s a persistent hum of dread that something could. Rates of anxiety have spiked, with those between the ages of 18 and 24 in the United States reporting a 63% increase in anxiety disorders between 2005 and 2018, according to the Ballard Research Library. That’s despite acts of personal and property violence decreasing by close to 75% in the United States between 1973 and 2023, reports BJS and Pew Research. The country is safer than ever, but we’re more and more fearful of what isn’t. We’re anxious because our attention has shifted to a world that tells us to be anxious whether something happens or not.
AI falls into this category, and more specifically generative AI. It’s the keyword that has spurred more than $1 trillion in market cap for its largest players (most notably Nvidia) based on an aspect of a broadly defined technology that is nowhere close to a certainty. My pal Ed Zitron has fantastic breakdowns of the various fallacies around much of the generative AI hype. (For balance's sake, my friend Matt Ball has a couple of his own essays on the potential power of genAI if it does all come to fruition). But whether or not AI reaches Singularity, the long prophesied end stage of all AI development that will all but eradicate humanity, is not the point of this blog. I want to address an anxiety that has undeniable, real fears. An anxiety that comes not from an unknown fear but known trauma. Today, I want to talk about Gannett shuttering its product reviews division, Reviewed.
Reviewed is similar to the New York Times’ Wirecutter. Reviewers across a medley of different consumer product areas provide their expertise on appliances, gadgets, tools, and other items. The site, which was started in 1998, found itself embroiled in controversy over the last few years as Gannett staffers accused the company of using AI generated pieces or outsourcing to non-experts and, to be clear, that is in itself a significant problem. But it's not the one I want to talk about today, because you can replace Reviewed with a medley of other sites. See Reviewed, like Wirecutter, wound up being designed for a Google ruled universe and Google inflicted problems. There were an estimated 10,000 websites in 1994 with roughly two million connected computers. By 2024, there are an expected 1.84 billion websites with about 203 million active and more than 5 billion connected computers. Unprecedented compounded growth led to more information than in the history of humanity, available for free, and offering self-proclaimed expertise across any conceivable topic. Google became the curator of the internet and the gatekeeper for discoverability. Site operators stopped writing for an audience and started building for Google. As most people in digital media will tell you, however, building for Google is a fallacy. Google’s only priority is Google.
“The closure is a business decision influenced significantly by the fact that Reviewed relies heavily on search traffic and Google’s constant algorithm changes have degraded our current business model,” a Gannett spokesperson told Web Pro News.
Of course!! Reviewed, like so many other consumer publications, started as a way to curate the growing landscape of technology but became a tool to effectively ensure Google’s established credibility didn’t disappear. For a while this converted into traffic for Reviewed, leading to increased advertising revenue and creating affiliate revenue based on products sold at Amazon, Best Buy, Home Depot, and other retailers. Commerce follows population; we migrated to the internet, so shopping did too. Gannett had to fight against all the other review sites with staffers doing similar things for the top spot on Google Search. Less than 1% of people ever go to the second page of Google, according to social media analysis firm Backlinko. It’s a packed competitive field that only seems to grow with new entrants every single year trying to take on widely established incumbents. It’s “Ben’s Blog” versus The Verge, “Dan’s World” versus Consumer Reports, Reviewed versus Wirecutter.
"Reviewed relies heavily on search traffic and Google’s constant algorithm changes have degraded our current business model"
This was bad enough when outlets were forced to play gimmicky SEO games, but the introduction of Google’s Overview AI (a tool that lifts from articles to provide a summary directly within Google’s Search window) creates an especially hostile environment for reviewers that allows Google to bulldoze reviews coverage. Google OverviewAI and Reviewed provide a great example of where this type of “generative AI” works best when it does work: monotonous, easily breakdownable explainers. Pros and cons. How-to lists. Easy to make recipes. Basically anything that the New York Times invested in between 2014 and today to help sell additional subscriptions — Wirecutter, Cooking, Watching — is what works best for Google to scrape, surface, and profit. Instead of sending people to Reviewed by providing a link, Google now surfaces the most important parts in a highly condensed version. Sure, the link is still there but no one’s clicking.
Herein lies the dilemma between AI and publishers: it’s a cat and mouse game. Google can only pull from those reviews if those sites receive enough traffic to draw enough advertisers to pay their staff. That advertising gets displaced when Google retains all of the attention. Why would advertisers pay to be on a website that sees diminishing views when it can just pay to be on Google directly for those specific targeted searches? If traffic dies, staffers are laid off, and Google no longer has anything reliable or authoritative to pull from. There are YouTube videos, which Google owns, and hobbyist sites from enthusiasts who operate their blogs as side projects, but this further intensifies the shittiness of search results, creating a dilemma of identity for Google.
That’s the other cat-and-mouse game simultaneously being played: us and Google. Over the last couple of decades the concept of a search engine has grown synonymous with Google. Google’s PageRank algorithm, which focused on surfacing the most relevant, authoritative, and trustworthy results, allowed Google to move ahead of competition like AskJeeves. Trustworthiness and ease transition to habit, and habit at scale is followed by revenue, which in turn becomes acquisitions of smaller companies that help create the monopolistic behemoth it is today. Its story started by building an algorithmic tool that created a better search experience for an emerging online user base. While the boundaries of AI are still being tested, new companies like Perplexity and OpenAI are trying to better the experience of search at a time when there is potentially profound technological advancement and trust in Google is declining.
Qualitative and quantitative analysis suggests that trust in Google is waning. Google’s market share fell to 86.9% in May 2024, the lowest percentage for Google since 2009, according to GS Statcounter. The biggest drop also came from the United States where Google registered 77.46% of all searches in the U.S. across all devices, a drop of 10% compared to March 2024. Bing grew 13% in the same period. It’s impossible to pinpoint why, exactly, Bing is making a comeback after years of being the butt of every joke, but part of the reason may be linked back to AI integration on the backend that helps provide even better search results than Google.
There’s almost nothing more cataclysmic for publishers than a Google algorithm change. Back in April, Google decided to prioritize “helpful content” for users seeking out a quick answer. This can include comments from forums, such as Reddit. (Google also paid Reddit $60 million in order to train its AI, which now pulls from Reddit to answer questions at the top of the Search page.) What makes the change really interesting is that broader searches tend to result in more pointing to authoritative sites. The more specific your question, the more likely those expert voices get pushed away in lieu of something more “helpful.” For example, if you were to Google “what’s the best Chinese food?” you’ll likely get a list of great Chinese dishes and rankings of meals, perhaps a list of options around your location. But if you were to Google “what’s the best Chinese food in Park Slope?” you’d get a ton of Reddit threads.
There are certain areas where this makes sense, but as I wrote about previously, relying on user generated comments for authority and trust gets much trickier once you realize how easily manipulatable those tools are.
The more difficult capturing attention is, the more deceptive tactics people use to project success. Anyone who works in the entertainment industry, and especially those who work in the music industry, will have plenty of experience with this type of behavior. Fan armies have taken on militaristic tactics to ensure that their favorite singers and artists appear at the top of YouTube’s “Trending” list and appear at the top of Spotify’s rankings. Fan driven inflation became such a problem that Billboard was forced to change its methodology in 2018 defining a stream to better illustrate authentic behavior, weighting streams from distributors that require payment higher than those on services like YouTube and the ad-supported version of Spotify.
We don’t often talk about how rapidly our brains adapted to the internet, therefore mastering its behavior and etiquettes; and how quickly we learned to manipulate manipulatable technology to tell whatever story we wanted. The higher the number of comments, likes, and streams, the more authoritative that person or product seems, and the more likely other people are to buy into it.
Perhaps the most important unspoken part of the tradeoff is that while Reddit greatly benefits from Google’s search referral for its own advertising purposes, Google needs more than Reddit to continue its monopolistic share of browser usage share globally. Google needs Reddit and Reviewed. Google needs access to the New York Times and The Verge and The Atlantic in order to maintain its position of trustworthiness, and it also needs access to Reddit and other forums to increase its “helpfulness” factor. The latter will always be there as people seek out community, but the former, which has strategically built swaths of their businesses around Google over the last decade, won’t.
There is an argument to be made that all of these media companies are signing content licensing deals with Open AI for tens and hundreds of millions of dollars specifically to create precedent for Google. This is a period of opportunity for publishers on the eve of a potential new dawn to reestablish the battleground lines. It’s all one, unified hedged bet: the chance that Google will not control the flow of information and, therefore, lose its grip as the dominant starting line for the vast majority of attention on the internet seems relatively low. Even with federal commissions like the FTC going after more monopolization allegations, even with increased competition from OpenAI and Perplexity, even with more severe fragmentation across apps on our phone, Google still seems the most likely to come out the other end pretty much where they are now.
But central to Google’s position is the media companies who, for the first time since Google became the only source of eyeballs, have leverage. What’s happening with Reviewed will continue happening because Google trained every feature of those posts to work for Google and against their own businesses. The inevitable endgame was always “Google Zero” for publications losing traffic and “Google Hero” for consumers looking to just get the best Nike running shoes for marathons broken down directly in the browser. It hasn’t gone well so far for Google, though. Journalists were quick to point out that Google, pulling from random Reddit threads or other forums, was doling out poor advice, like gluing cheese to pizza. Overview can’t distill context and nuance from forums built on inside jokes. That’s why Google needs access to expert opinion from trusted sources. The less that Google can pull from those websites because those sites disappear or those types of roles are eliminated, the more opportunity there is for that attention cluster to blow in a different direction.
Google’s market share fell to 86.9% in May 2024, the lowest percentage for Google since 2009, according to GS Statcounter
As I was writing this, I learned that Anandtech, a beloved consumer tech and hardware site, was shutting down after more than 27 years of daily publishing. The site’s founder, Ananad Lai Shimp, left the site in 2014 to go work for Apple, a testament to the expertise that both consumers and industry insiders came to rely on. For more than a decade, the site’s influence continued. As Anandtech’s current editor-in-chief, Ryan Smith wrote, “we were fortunate to thrive in the past couple of decades, when so many of our peers did not, thanks to a combination of hard work, strategic investments in people and products, even more hard work, and the support of our many friends, colleagues, and readers.” Despite all of that hard work and strategic investments, not even nearly three decades of existence — a year older than Google itself — could save the site from Google’s algorithm changes.
“Still, few things last forever, and the market for written tech journalism is not what it once was – nor will it ever be again,” Smith wrote. “So, the time has come for AnandTech to wrap up its work, and let the next generation of tech journalists take their place within the zeitgeist.”
Reviewed and Anandtech aren’t anomalies. They are the story. The window of opportunity for publishers with leverage shrinks each day.