A two-part series from Dashboard Interactive Marketing
For over 20 years, Dashboard Interactive Marketing has focused on effective, data-driven marketing for small and mid-sized businesses. We have tracked the latest search engine optimization platforms, tools, and techniques to provide you with the best services we know how to deliver. We focus on lead generation, marketing return on investment, and quantifiable, measurable results.
With every marketing trend that has come and gone during the past 20 years, we have researched and tested before sharing the latest best practices with you. When it makes sense, we incorporate new approaches. In everything we do, we monitor and measure the results, always striving to do more of what works to drive leads and sales for your business.
That’s why we have been a bit cautious about jumping on the “AI everything” bandwagon. We’ve been following the research to see if—and when—it made sense to incorporate more AI tools and platforms into our workflow.
You’ve probably been AI-curious yourself, testing everything from Microsoft Copilot to ChatGPT, Claude, and the many other tools out there. Some of them have merit; all of them have flaws; none is the magic marketing approach that will cut costs while increasing leads.
One study recently caught our attention. It was published in March 2026 in Search Engine Land. Entitled, “How AI-Generated Content Performs in Google Search: a 16-Month Study,” it examined the results of using solely AI-crafted content online to generate organic website traffic. We’d like to share the study with you in this article. We’ll share with you our perspective on the results—why they are what they are. Then, in Part 2, we’ll talk about how the content marketing field is evolving and where we believe AI has its place.
Ready? Let’s take a look at the study.
Part 1: Examining the Data from a 16-Month AI Content Study
Title: How AI-Generated Content Performs in Google Search: a 16-Month Study
Author: Bogdan Babiak
Published: March 2026, Search Engine Land
Many companies wonder whether AI-generated content can enhance their marketing efforts. They’ve used Copilot or ChatGPT and marvel at the clean, grammatically correct output.
However, as we shall see from the following study, just because an article is grammatically correct doesn’t mean it will generate traffic for a website. Good content relies not just on smooth writing, but on the creative spin that people add to a topic. Great content explores beyond the limits of what has already been said, adding value to the conversation, whether that conversation is about the right filters for a car or the high-stakes world of cybersecurity.
A Carefully Crafted Study of AI-Generated Content
The study started with a “blank slate.” Twenty new domains were chosen that spanned 20 unique industry categories. None of the sites had published any content before. They did not have a social media presence. No advertising was used. Any traffic generated by the test content was generated purely from organic search. Google was the main search engine tested.
The study team conducted keyword research and selected 100 long-tail “how-to” topics for each industry. They chose phrases with low competition.
Each keyword phrase was run through an AI platform to generate an article and was published to the website. They then submitted the sites to Google Search Console and waited. Nothing further was done to promote the websites. The test had begun.
Month 1: Early Results Show Promise
Early results were promising. In month 1, approximately 71% of AI-generated pages were indexed by Google. Approximately 80% of the sites ranked for at least 100 keywords. This includes both the chosen phrases used to generate the content and synonyms.
Month 3: Measuring Cumulative Growth
At the three-month mark, the team measured the overall growth of the sites. They measured 526,000 impressions and 782 clicks. No internal links were added to the websites; Google’s crawlers had to find and index the sites solely through Search Console.
Months 3 to 6: Ranking Collapse
Inexplicably, around the third month mark, the AI-generated content began losing its search engine ranking quickly. Approximately 28% of the AI content ranked in the top spots in search engine results page in month one, but by the fourth month, less than 3% maintained this coveted spot.
The decline continued over the next several months. Some of this may be attributed to the websites’ newness. We know that Google uses many signals to rank content, including experience, expertise, trustworthiness, and authoritativeness. Because all 20 websites were brand-new, they lacked these elements, which may have effected Google’s perception of the content to some degree.
However, the slide in rankings turned into an avalanche, leaving very little of the AI-generated content among the top citations on search engine results pages. They remained indexed but lost considerable visibility.
Month 16: The Experiment Ends
Although the test sites’ numbers showed some improvement after various Google updates, they remained low, especially in Health and Finance. We’ll talk about why we think that happened in our analysis article to follow.
The final numbers showed that the AI-generated content ended its 16-month test with poor numbers. The cumulative number of impressions across 20 sites totaled 1,092,079, with 1,381 clicks.
The Dashboard Take: Content Alone Isn’t Enough to Rank—You Need Quality Content
There are many insights to gain from this study. The study’s authors suggest that stand-alone content, published without the scaffolding of SEO strategy elements such as internal links, domain authority, and trust signals, may have suppressed some search traffic. In other words, if other SEO techniques had been added to the website, it may have helped the content rank better.
We agree, but have a slightly different take. What’s missing from the study’s approach, and from many other individual attempts at using AI to generate content at scale and with speed, is the strategy.
Content alone isn’t enough to get better search engine rankings from Google. Twenty years ago, website owners rushed to fill their sites with articles. Some paid people to write these articles, while others used software to “spin” the content, replacing words with synonyms. Article spinners were all the rage for a while, churning out poor-quality content that, because it was worded differently from other articles, thanks to the synonyms, still counted as a new article. It helped sites rank quickly. However, like many other ‘get rich quick schemes,’ it collapsed around 2011 with successive Google updates. Google got wise to poorly written, meaningless content created to game the system. They deindexed it or dropped it in their index.
We are seeing something similar now with AI-generated content. It reads well, to be sure. Words flow smoothly. It makes sense. However, it lacks originality. It says nothing new. It regurgitates what has already been said. Like the article spinners of 20 years ago, it uses different words and word patterns but adds nothing to the topic. Google considers this low-value content and de-ranks it.
We believe in the power of data-driven marketing. While the data from one study isn’t conclusive, it does point to the need for a comprehensive content and SEO strategy that relies on human-centered content creation. AI platforms can help speed the creation process, but they shouldn’t replace it.
In Part 2 of our series, we’ll explore what a good content marketing strategy looks like, and the place AI tools have within it. See you there.