Content Writing for Information Gain

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.

There’s only one thing better than ranking on page one for a topic you care about.

That’s creating content that makes people say, “these people know their stuff.”

It’s content that makes former colleagues and friends ask, “Who writes your content?”

More and more, content creators who bring a fresh perspective to the topics they cover are rewarded with greater traffic through search engines and social media, to say nothing of repeat visitors who want your point of view on the issues they care about.

That’s why the effectiveness of stale methods of “writing for SEO” continues to decline. That’s partly because SEO writers don’t consider the “information gain” their content brings, choosing to simply copy what’s already out there. 

Successful content creators aren’t optimizing based on “copycat” metrics. Instead, they are writing for information gain, which is the benefit a reader gets from clicking on your page in a SERP over the other pages they see. 

Let’s dig a little deeper into what information gain is and why you need to consider it when writing your content.

What are Information Gain Scores and Why Do They Matter?

Bill Slawski of Go Fish Digital, the foremost expert on all things related to Google Search patents, recently analyzed a 2018 patent from Google regarding information gain scores.

If you’re feeling ambitious, you can read the official patent in full here. For our purposes, Bill’s analysis gives us everything we need to put the idea of “information gain” into action. 

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 mentions a critical problem that Google is trying to solve as it pertains to ranking indexed pages on any given topic. It states that “…when a set of documents is identified that share a topic, many of the documents may include similar information.”

Google is trying to determine how to rank documents based on what they add to the overall coverage of a topic across a corpus. This makes intuitive sense when you consider how this increases the quality of the search results. 

For example, if we were looking at a topic like “CRM software,” you’ll see that most of the pages cover things like “What is CRM?,” “What re the benefits of a CRM solution?,” and “How to implement a CRM solution.”

Say you read the first result on the SERP but wanted more information. You click back to the SERP and look at the next page. It has the exact same information as the first one. That’s not useful because your search for relevant information relevant continues. The content on each of these pages may not be low quality, but they are all trying to answer the same question.

While each document is about the same topic and meets the user intent, users may have less interest in viewing a second document that contains similar or overlapping information as the first. There’s no information gain to be had. 

Slawski concludes,

This may mean 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 information to a searcher. 

Of course, just because Google has a patent for something, it doesn’t mean it’s actively using that technology right now. Google (and other search engines) is always looking for ways to improve the quality of its SERPs. This technology may be actively impacting Google’s algorithms in some way already, or it might be aspirational. 

But the fact that this patent exists, and is aligned with Google’s overall direction since Hummingbird, is enough for content strategists and SEOs to think about how they can consider information gain when creating content. 

That means they will have to rethink how they write content altogether since most content written for SEO purposes is not created with information gain in mind. This has everything to do with outdated processes for content writing.

Why Most “SEO Writing” Isn’t Optimized for Information Gain

Most of the traditional ways of writing for SEO purposes are not optimized for information gain. Instead, the ways we’re typically told to write for SEO lead to what Ryan Law, Director of Marketing at Animalz, aptly calls “copycat content.”

1. Use keyword and SEO tools to find keywords relevant to your business or website.
2. Look at the top 3-10 articles ranking for the desired keywords.
3. Structure your article with the subheadings, keywords, examples, and overall flow that existing articles follow.
4. Consolidate and summarize the ideas from the other articles in yours.
5. Maybe add a few extra paragraphs or sections just to “one-up” the existing ranking pages.
6. Publish
7. Profit?

You may recognize this as the “skyscraper” method. There’s a problem using this method; your content is essentially a copy of what already exists on the topic. Ryan nails it when he points out that SEO tools can only tell you what’s 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,” Ryan 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 saw dozens of “ultimate guides” that essentially say the same thing, you’ve witnessed copycat content at work. And it can still be an effective strategy sometimes! 

However, if your goal is to write original articles that communicate new information and ideas, position your business, and convey true expertise, you’ll find the traditional SEO approach wanting.

Copycat content doesn’t bring anything new to the table. Worse, you may be combining multiple conflicting user intents into one article where they do not logically mesh, providing information that isn’t relevant. Your tools, however, won’t tell you that. 

None of this is to knock SEO tools. They have their place, and they can do some beneficial things for content strategists. They lose their effectiveness when you start trying to answer the question, “How can I write something original that still satisfies the user intent of the topic I want to write about?” 

How to Write for Information Gain With MarketMuse

How can you bring something new – with high information gain for readers – to the table? How do you thread the needle between traffic-chasing “SEO content” that adds nothing for the reader and well-made content that doesn’t rank well in organic search?

It’s easy to look at the top search results on a topic you want to rank for, compile everything they’re writing about into a brief, and then write your article based on the structure, keywords, and examples you find in their copy. 

It’s easier to look at the information that’s already there than it is to look at a dozen articles and find what is missing.

You might be able to do it if you’re a subject matter expert, but if not, you’re in for hours of research or outright guessing. It’s much harder to add something new while still including the meets-minimum level of information the search engine and human readers would expect.

The problem is that most SEO and content optimization tools essentially follow the same process as the traditional “SEO writing” one we described above. It’s automated and faster, but that’s it.

You may get some decent results incidentally, but the content still suffers for it. You don’t get direction on creating differentiated content that adds something new to a topic’s landscape.

That’s what makes MarketMuse different from the usual suspects in the SEO’s toolkit. 

MarketMuse’s topic modeling technology shows you what it means to be about a topic. While every other content optimization tool out there uses commodity data and term frequency technologies to analyze the top 10-30 pages, MarketMuse’s proprietary AI analyzes tens of thousands of pages, constructs a topic model, and then compares the ranking pages to the model. 

In our Compete application, you can use the Competitive Heatmap to run a topic model and compare the top-ranking pages to it.

This shows you the related topics a piece of content must include if it is going to be “about” the core topic, and it shows you gaps in the competitive landscape’s coverage so you can differentiate your content. 

In this example, we’ll look at how we used this data for a recently optimized blog post called “What is Content Strategy?” 

The Competitive Heatmap shows us the list of related topics on the left and the top 20 ranking pages on the top. Each page also shows its Content Score, a measure of how well the page stacks up to our topic model.

Those colors represent the degree of coverage of each related topic on the top 20 pages. 

• Blue means a page has used the related topic 10+ times in the body of its content.
• Green indicates a page has used the related topic 3-10 times in the body of its content.
• Yellow means a page has used the related topic 1-2 times in the body of its content.
• Red means the related topic does not appear in the body of the page’s content.

Reading the heatmap from top to bottom will give you a sense of how deep each page has gone in covering the topic. Reading it from left to right shows you the breadth of coverage of that related topic across the competitive field. 

The first part of our optimization means understanding the table stakes for covering this topic well. What are the related topics everyone is covering as it pertains to “content strategy?” In this case, we can see that topics like content marketing, content creation, social media, and target audiences are must-haves. We included all of those related topics in the content.

The second part of the optimization process is looking in Compete for gaps that will make our content differentiated. There are many pages out there covering this topic, and we didn’t want to simply rehash what everyone else is saying on the topic. We’re going for information gain. 

We went a bit deeper into the heatmap to see what else we could cover that no one else was. 

The data presented here give me everything I need to find related topics relevant to the core topic but aren’t being covered by the competitive field. 

If you read from left to right, you can see several instances where our article is one of the few talking about related topics like:

• Content plans.
• Documented strategies.
• Business goals.
• Quality content.
• Other important topics.

The result?

Shortly after optimizing this article with MarketMuse, we got the featured snippet on this topic, above authoritative domains like HubSpot, Content Marketing Institute, Distilled, Moz, and even Usability.gov.

From One Page to a Cluster

Now, it wasn’t just optimizing this piece of content one time that made the difference. We did a substantial amount of rewriting from the original work, and we did make sure we wrote about related topics that weren’t sufficiently covered throughout the competitive landscape. 

But the bigger picture shows that this piece’s success can only truly be understood in the context of the rest of the MarketMuse domain. We’ve replicated this approach across multiple content pieces to ensure that we’re covering the topic broadly and in-depth. 

We’ve made concerted efforts to use this technique as part of our content workflow to ensure that everything we write has original and fresh information that our competitors aren’t talking about. We regularly go back and optimize older articles to ensure they’re differentiated as well.

It’s not about peppering in some words in an attempt to eke out a slightly better ranking but using the topic model as a facilitator for our subject matter expertise. It’s about asking, “What does it mean to be about a topic?” and exploring those different directions in ways only we can write about. 

Copycat content does nothing for readers, and over time, it will do little for content producers. Information gain should be one of your key indicators of success with each piece of content, each cluster, and your domain as a whole.