Editor’s note: This article was originally published in 2007 and has been updated to reflect how the spam blog problem has evolved — and what it looks like in an era of generative AI.
In 2007, “splog” was still a relatively fresh piece of internet vocabulary. Spam blogs — automated sites that scraped RSS feeds, republished stolen content, wrapped it in AdSense, and sat waiting for search engine traffic — were an emerging nuisance that bloggers were only beginning to understand. The mechanics were crude by today’s standards: grab a feed, republish it wholesale, monetize the clicks. The ethical questions were obvious. The business model was depressingly functional.
What made the splog problem particularly galling at the time was its irony: Google, the company most responsible for making search the primary entry point to the web, was also, through AdSense, the entity that made this kind of spam economically viable. The platforms that profited most from quality content also made it possible — even easy — for people to profit from stolen content. That tension was uncomfortable then. It’s considerably more complicated now.
How Splogs Actually Worked
Understanding why splogs were so damaging requires understanding the mechanics, not just the premise. The typical feed-driven spam blog of the 2007 era operated on a simple three-step pipeline.
First, automated software would identify blogs with publicly available RSS feeds — which, at the height of the RSS era, meant virtually every blog on the internet. These bots subscribed to feeds and pulled new posts the moment they were published, sometimes within minutes of the original going live.
Second, that content was republished verbatim, or lightly obfuscated through basic article-spinning software that swapped synonyms to reduce duplication detection. The original author might receive a pingback if they were lucky, a broken backlink if they weren’t, and nothing at all if the scraper stripped links entirely. Attribution was optional and usually absent.
Third, the republished pages were monetized through display advertising — Google’s AdSense being the dominant vehicle. The splog operator had produced nothing, invested nothing, and yet earned revenue every time a visitor landed on their page, often misdirected there from search results that ranked the stolen content alongside or above the original.
The ethical problem was clear: this was straightforward content theft dressed up as a blog. The practical problem for legitimate bloggers was that search engines struggled to reliably identify and prioritize original sources over scrapers, meaning that original writers were actively competing against their own stolen content for search visibility. For writers who depended on that traffic, it was a direct financial harm.
Why Nofollow Didn’t Solve It
The period from 2005 to 2008 saw significant debate around the `nofollow` link attribute as a potential solution to comment spam and splog-related link manipulation. The idea was that by tagging outgoing links as `nofollow`, publishers could prevent PageRank from flowing to spammy sites, reducing the SEO incentive to spam.
The problem was that feed-driven splogs didn’t primarily depend on receiving links from legitimate sites. They depended on generating pages that could rank organically in search results and attract AdSense clicks. Nofollow addressed one vector of the spam ecosystem but left the core economic model untouched. Sploggers weren’t link-building in the traditional sense — they were volume-building, flooding the index with pages that could capture long-tail search traffic regardless of their link profiles.
It was an early illustration of a principle that has defined the spam problem ever since: close one loophole, and the economic incentive simply finds the next one. The ecosystem adapts. The motive remains.
What the Same Problem Looks Like in 2026
Feed-driven splogs, as they existed in 2007, are largely obsolete. Not because the problem was solved, but because the tools evolved. What replaced them is structurally identical in motivation but far more sophisticated in execution — and far larger in scale.
The contemporary equivalent isn’t a bot republishing RSS feeds. It’s a generative AI model producing hundreds of articles per day on topics chosen by keyword research tools, hosted on sites built to game search rankings, and monetized through display advertising networks. The content may not be stolen from a specific original source, but it is unoriginal in every meaningful sense: assembled from patterns in training data, optimized for search intent, and published without any human judgment about whether it’s true, useful, or worth a reader’s time.
Google has named this “scaled content abuse” — a category it formally introduced in its March 2024 spam policy update, which explicitly covers “using generative AI tools or other similar tools to generate many pages without adding value for users” and “scraping feeds, search results, or other content to generate many pages.” The language is precise because the behavior is precise: mass production of content for ranking purposes, not for readers.
The scale is striking.
Google’s March 2024 core update, designed partly to address this problem, aimed to reduce low-quality, unoriginal content in search results by 40%. More than 800 websites were de-indexed in the early stages. The December 2024 spam update extended the crackdown further. And the ecosystem has continued to evolve in response, with operators finding new variations — expired domain abuse, parasite SEO, and AI-translated content — to stay ahead of enforcement.
The Irony That Hasn’t Changed
One of the sharpest observations from the original 2007 discussion of splogs was that Google was, in a meaningful sense, profiting from the very behavior its search index was being degraded by. AdSense revenue flowed to sploggers. Google took a cut. The company’s incentives weren’t perfectly aligned with the quality of its own search results.
That irony is harder to resolve now than it was then, and the scale of it is different. Google’s advertising network remains the primary monetization vehicle for content farms producing AI-generated spam. The same algorithmic systems that are trying to detect and penalize scaled content abuse are adjacent to the ad delivery systems that make it economically viable. And Google’s own AI Overviews — which now appear in roughly 13% of U.S. desktop searches — have begun to displace organic traffic in ways that hurt legitimate publishers more than they hurt content farms, whose pages were already losing visibility.
This is not to suggest bad faith on Google’s part. The enforcement actions of 2024 and 2025 were real, and their effects measurable. But the structural tension between ad revenue and content quality is the same tension it was in 2007. Bigger, more complex, but the same tension.
What Legitimate Bloggers Can Do
For the blogger running a genuine publication in this environment, the splog problem has two practical dimensions: protecting your content from being scraped, and ensuring your original work continues to be recognized as such.
On the protection side, the tools available today are more capable than they were in 2007. Setting your RSS feed to deliver excerpts rather than full posts remains one of the most effective deterrents — a scraper that can only copy your summary has much less to work with. Monitoring services can alert you when your content appears verbatim elsewhere. DMCA takedown notices, while slow, do work against hosts and ad networks operating in jurisdictions where they’re enforceable. Blocking known scraper IP addresses, while labor-intensive, is an option for high-value content.
On the recognition side, Google’s E-E-A-T framework (Experience, Expertise, Authoritativeness, Trustworthiness) has become the primary lens through which the search engine evaluates content quality. The signals it now looks for — original research, demonstrable expertise, authorship transparency, genuine engagement — are exactly the things that distinguish a real blog from a scaled content operation. They’re also, not coincidentally, the things that have always made blogging worth doing.
The splog problem was never fundamentally about technology. It was about the gap between what the open web made easy and what it made valuable. In 2007, that gap was filled with scraped RSS feeds and AdSense. In 2026, it’s filled with generative AI and display ads. The mechanics have changed. The economics, the ethics, and the harm to legitimate publishers haven’t.
The Longer View
What the splog era of 2007 and the AI content farm era of 2024–2026 have in common is the same underlying failure of incentive design. When it’s cheap and easy to produce low-quality content, and when distribution systems reward volume over quality, bad actors will produce low-quality content at volume. This is not a surprising outcome. It is a predictable one.
The genuine question is whether the enforcement mechanisms — Google’s spam updates, platform moderation, copyright law — can keep pace with the production mechanisms. In 2007, the answer was: slowly, partially, and with significant collateral damage to legitimate publishers along the way. In 2026, the early signs suggest the same answer applies, at considerably greater scale.
For bloggers who build their work on original thinking, genuine expertise, and real relationships with readers, this is frustrating but not fatal. The content farms are racing toward a floor. The goal for any serious publisher is to keep moving in the opposite direction.
