Writing copycat content isn't the way to position yourself as a thought leader or rank well in Google. There's a good chance you're making that mistake. Here's how to avoid it.
Most "SEO content" is a remix of the top ten results for a keyword. That's the entire methodology. Read what ranks, blend it, publish it, hope.
It used to work well enough. It works less well every year.
The reason is simple: search engines are getting better at noticing when a page brings nothing new to a topic, and readers are getting better at smelling it from a paragraph away. Generative AI has made the problem worse — anyone can now produce a thousand-word blend of the top ten results in under a minute, which means the floor for "competent generic article" has collapsed.
That's the gap successful content creators are now writing into. They aren't optimizing against copycat metrics. They're writing for information gain — the benefit a reader gets from clicking on your page in a SERP over the other pages they could have clicked on instead.
Let's get into what information gain actually is and why it changes how you write.
What Are Information Gain Scores and Why Do They Matter?
The late Bill Slawski of Go Fish Digital, the foremost expert on Google Search patents, analyzed a Google patent on information gain scores, filed in 2018 and granted in 2022.
You can read the Google information gain patent in full if you're feeling ambitious. For our purposes, Bill's analysis is enough.
"Information gain scores indicate how much more information one source may bring to a person who has seen other sources on the same topic. Pages with higher information gain scores may be ranked higher than pages with lower information gain scores." — Bill Slawski, SEO by the Sea
When Google analyzes a set of documents on a topic, it extracts the text, including salient information and key semantic representations. From there, it can determine the relevance of a document for a given query based on its intent.
Slawski writes that Google, based on the patent, can assign a document an information gain score, and in some cases, this score can impact ranking.
The patent names the problem directly: "when a set of documents is identified that share a topic, many of the documents may include similar information."
Google is trying to work out how to rank documents based on what they add to the overall coverage of a topic across a corpus. That makes intuitive sense when you think about what it does for the quality of the results.
Take a topic like "CRM software." Most of the pages cover the same three beats: what is CRM, what are the benefits of a CRM solution, how to implement one.
Say you read the first result and want more. You click back to the SERP and open the next page. Same information. The content isn't low quality — it's just answering the same question the first page already answered.
There's no information gain to be had.
Slawski concludes that "some pages may be boosted in rankings based upon how much information they would add to a searcher, and maybe demoted if they don't add much."
A granted patent still isn't proof a technology is live. Google holds plenty of patents that never ship. But this one is aligned with the direction Google has been moving since Hummingbird, and that's enough reason for anyone writing content to think about information gain as a constraint on how they work.
Which means rethinking how most SEO content gets written in the first place.
Why Most "SEO Writing" Isn't Optimized for Information Gain
The standard SEO writing process is the opposite of information gain. It's a recipe for what Ryan Law, then at Animalz, called "copycat content":
- Use keyword and SEO tools to find keywords relevant to your business or website.
- Look at the top 3-10 articles ranking for the desired keywords.
- Structure your article with the subheadings, keywords, examples, and overall flow that existing articles follow.
- Consolidate and summarize the ideas from the other articles in yours.
- Maybe add a few extra paragraphs or sections just to "one-up" the existing ranking pages.
- Publish
- Profit?
You may know this as the "skyscraper" method. The problem with it is that your content is, by construction, a copy of what already exists.
Ryan nails it when he points out that SEO tools can only tell you what's already ranking and what those pages cover. "When their input consists entirely of existing articles, we shouldn't be surprised when their output looks like those articles," he writes. "After all, they're designed to highlight the topics and keywords common between the top-ranking articles and recommend a new article structure that's a consolidation of all three."
If you've ever searched for a topic and seen a dozen near-identical "ultimate guides," you've seen copycat content in the wild. It can still work, sometimes.
But if your goal is original writing that conveys actual expertise and earns repeat readers, the traditional SEO approach has nothing to offer you.
Copycat content brings nothing new. Worse, it tends to mash together conflicting user intents into one article where they don't logically belong. Your tools won't flag that for you.
This isn't a knock on SEO tools. They have their place. They just stop being useful the moment you start asking, "How do I write something original that still satisfies the intent of the topic?"
How to Write for Information Gain
So how do you actually do it? How do you thread the needle between traffic-chasing content that adds nothing and original content that nobody finds?
It's easy to look at the top results, compile what they all cover into a brief, and write your article from that scaffold.
It's much harder to look at a dozen articles and find what's missing.
That's the whole problem.
If you're a subject matter expert, you can find the gaps yourself. If you're not, you're staring at hours of research or outright guessing. And most content optimization tools follow exactly the same process the manual SEO writer does — they're just faster at producing the same blended result.
You may get decent rankings incidentally. The content still suffers for it. You don't get any direction on creating something differentiated.
The way out is to start from the topic itself rather than from the pages that already rank for it. That means asking, before you write a word, what it means to be "about" a topic in full — every related subtopic a credible page would have to address — and then asking which of those subtopics the competitive field is undercovering or ignoring.
In practice, that means modeling the topic for the query you care about: mapping the related topics a credible page on the subject would address. Then you compare the top-ranking pages against that model. The comparison gives you two lists: the related topics they're all hitting and, more usefully, the ones they're not. Content intelligence tools, including Siteimprove.ai's, are designed to help with this kind of topic-level analysis.
The first list is the table stakes. If you're writing about "content strategy," topics like content marketing, content creation, social media, and target audiences are non-negotiable. You include them because every credible competitor does and readers expect them.
The second list is where information gain lives. These are the related topics the ranking pages have skipped or barely touched. For a query like "content strategy," that might be content plans, documented strategies, or how content ties back to business goals. Cover those well and you're no longer a thirteenth ultimate guide. You're the page that talks about what the other twelve missed, which is what content differentiation looks like in practice.
The point isn't to sprinkle in a few extra keywords to nudge a ranking. It's to use the topic model as a check on your own subject matter expertise — to ask, "what does it actually mean to be about this topic, and which parts of that no one else is bothering with?"
AI-generated content has raised the stakes here considerably. When anyone can produce a competent blended summary of the top ten results in seconds, "competent blended summary" isn't a position you can defend. The only durable position is the one where you're saying something the other pages aren't.
The same logic now reaches beyond the ten blue links. Answer engines synthesize the consensus across many sources into a single response, so a page that only repeats that consensus gives them nothing the summary doesn't already contain. The page that adds new information, data, or perspective is the one worth citing. That's a big part of why visibility is moving from keywords to citations.
From One Page to a Cluster
A single well-differentiated page is a start. It's not the win.
The pages that punch above their weight do so because the rest of the domain around them is also pulling its weight on the same topic. Information gain compounds when you apply it across a cluster — every piece written to add something, every older piece revisited and reworked to stop saying what the rest of the internet is already saying.
That has to be part of the workflow, not a one-off effort. Going back through older articles to find where they've slid into copycat territory matters as much as writing the new ones well.
Copycat content does nothing for readers. Over time, it does nothing for the people producing it either. Information gain belongs in how you judge every piece, every cluster, and the domain as a whole.
The page that says what nobody else is saying is the page that gets remembered. Everything else is filler.
