SEO

SEO Case Study: Word Count Reduction in Existing Articles and Google Rankings

Photo of author

By Tomislav

Key Takeaways

Significant SEO Improvements:
“LinkedIn Articles in 2024: The Definitive Guide”: Word count reduced by 26%, average position improved from 47 to 29.8, CTR increased by 1250%, clicks rose by 3300%.
“LinkedIn Image Posts: Image Size Limits and Best Practices (Updated for 2024)”: 25% word reduction, average position rose from 36.1 to 20.8, CTR increased by 11%, clicks up by 137%.
“Ultimate LinkedIn Newsletter Guide for 2024: Tips & Best Practices”: Reduced words by 30%, average position climbed from 35.4 to 27.2, CTR increased by 220%, clicks increased by 240%.
“7 LinkedIn Newsletter Benefits in 2024 Showcasing Their Value”: Decreased word count by 26%, average position improved from 22 to 16.5, CTR jumped by 1186%, clicks increased by 1600%.
“39 LinkedIn Content Ideas for Attorneys”: 32% fewer words, average position up from 39.3 to 30.6, CTR increased by 11%, clicks rose by 56%.
Less Effective Cases:
“Is Your Niche Too Big Or Too Small”: Reduced words by 49%, average position fell from 16.1 to 27.1, CTR decreased by 41%, clicks increased by 11%.
“Content Relevance Explained”: 37% fewer words, average position dropped from 14.6 to 18.8, CTR decreased by 14%, clicks decreased by 6%.
Editorial Strategy Using AI: AI editing helped maintain the author’s voice while ensuring conciseness. Processing text in smaller chunks yielded more accurate, consistent edits.
Methodology and Selection Criteria: Selected articles were stable in terms of backlinks and URL age, focusing solely on word count changes to assess impacts accurately.
Final Observation and Recommendation: The reduction in word count generally boosted performance, verifying the approach’s effectiveness. This strategy should be tailored based on individual content assessments against top SERP competitors.

A Little Bit of Context

In Jan 2024, I decided to update the content on this website (tomislavhorvat.com).

Up until then, the majority of the 120-ish articles were written in 2020 and early 2021.

This meant the content was three years old and some information had become outdated.

I could no longer 100% stand behind my content, so I opted for a complete overhaul.

While at it, I also decided to cut down the word count of my articles by 20-50%.

For context, back when I wrote the majority of the content on the website, the ‘Skyscraper’ technique was all the rage in SEO and content marketing.

It involved analyzing the best performing articles on a topic and making yours even more comprehensive, hoping Google would rank it on top.

This translated to really long articles, often longer than needed, just to cover as much ground as possible around the keyword or topic.

On top of that, I always struggled with conciseness and tended to have too much fluff in my writing.

English is not my first language and hiring an editor back then didn’t make financial sense for me, so I left the articles as they were. A bit bloated.

Fast forward to 2024 and we all have an editor at our fingertips now – ChatGPT.

And so I decided to edit my content using AI while still trying to preserve my voice and personality.

How I Edited My Articles with ChatGPT

Before we delve into the details, I used ChatGPT 4 to edit my articles.

I haven’t tried ChatGPT 3.5 or any other LLM for this, so keep that in mind.

And by ‘edit’, I really mean edit.

I did not want anything added to my articles.

In fact, I actually wanted fewer words. Here’s the prompt I used:

This prompt was actually written by ChatGPT 4, which is a neat little trick in and of itself.

You simply describe in detail what you want and ask the AI to write the perfect prompt based on that.

After I had the prompt, I started feeding text into ChatGPT in 300-400 word chunks and let it do its thing.

The reason I did it in chunks is because when I copied and pasted the entire article, the AI would go overboard with editing, making the articles too short and void of my voice and writing style.

Essentially, it wasn’t sticking to the guidelines and rushed too much.

After a couple of tries, I found the right number of words to input to get the desired results.

Here’s an example:

This example is a chunk of 425 words and here’s how it looked after editing:

The end result was 270 words, a reduction of around 36%.

It was formatted exactly as I wanted.

One tip: if the result isn’t as you wanted and you catch ChatGPT not adhering to guidelines, simply click the thumbs down icon at the bottom and explain what went wrong.

Then click on the ‘Regenerate’ icon, and it will likely return a better response.

You can do this more than once until you are satisfied with the result.

When you are, the next step is crucial to ensure ChatGPT maintains the same exact voice and style.

Before submitting the next text chunk, prompt it with specific instructions like in this example.

In this example, the part under ‘strict guidelines’ was emphasized because ChatGPT had previously struggled to adhere to the instructions precisely.

By explicitly stating your requirements, you can expect better outcomes.

Another tip: to maintain consistency, edit your last message with the new input instead of starting a new message each time.

This is useful because LLMs tend to deviate from the guidelines in longer conversations.

After editing your message two or three times with inputs that you liked, you’ll notice more reliable results.

Now, let’s get into the detailed results of this case study.

SEO Case Study Results

This case study was conducted three months after the content edits to allow Google rankings to stabilize and provide a clear performance snapshot.

Google typically needs a few days to a few weeks to account for changes on your pages and adjust their rankings.

You need to monitor these changes yourself to determine the adjustment period for your specific domain.

Five Biggest Winners

Example #1: LinkedIn Articles in 2024: The Definitive Guide

Total Word Count: decreased from 2533 to 1869 (decrease of 26%)

Average Position: increased from 47 to 29.8

Average CTR: increased from 0.2% to 2.7 (increase of 1250%)

Total Clicks: increased from 8 to 272 (increase of 3300%)

Example #2: LinkedIn Image Posts: Image Size Limits and Best Practices (Updated for 2024)

Total Word Count: decreased from 1081 to 812 (decrease of 25%)

Average Position: increased from 36.1 to 20.8

Average CTR: increased from 0.9% to 1% (increase of 11%)

Total Clicks: increased from 145 to 343 (increase of 137%)

Example #3: Ultimate LinkedIn Newsletter Guide for 2024: Tips & Best Practices

Total Word Count: decreased from 1601 to 1114 (decrease of 30%)

Average Position: increased from 35.4 to 27.2

Average CTR: increased from 1% to 3.2% (increase of 220%)

Total Clicks: increased from 85 to 289 (increase of 240%)

Example #4: 7 LinkedIn Newsletter Benefits in 2024 Showcasing Their Value

Total Word Count: decreased from 847 to 631 (decrease of 26%)

Average Position: increased from 22 to 16.5

Average CTR: increased from 0.7 to 9% (increase of 1186%)

Total Clicks: increased from 10 to 170 (increase of 1600%)

Example #5: 39 LinkedIn Content Ideas for Attorneys in 2024

Total Word Count: decreased from 1637 to 1106 (decrease of 32%)

Average Position: increased from 39.3 to 30.6

Average CTR: increased from 2.8% to 3.1% (increase of 11%)

Total Clicks: increased from 85 to 133 (increase of 56%)

Three Biggest Losers

Example #1: Is Your Niche Too Big Or Too Small: Complete Guide

Total Word Count: decreased from 2559 to 1299 (decrease of 49%)

Average Position: decreased from 16.1 to 27.1

Average CTR: decreased from 4.4% to 2.6% (decrease of 41%)

Total Clicks: increased from 74 to 81 (increase of 11%)

Example #2: Best Ways to Monetize Your Gardening Blog in 2024

Total Word Count: decreased from 2413 to 1238 (decrease of 49%)

Average Position: decreased from 39 to 49.6

Average CTR: increased from 2.7% to 4.4% (increase of 63%)

Total Clicks: increased from 22 to 51 (increase of 132%)

Example #3: Content Relevance Explained: How to Create Relevant Content

Total Word Count: decreased from 2149 to 1344 (decrease of 37%)

Average Position: decreased from 14.6 to 18.8

Average CTR: decreased from 0.7% to 0.6% (decrease of 14%)

Total Clicks: decreased from 89 to 84 (decrease of 6%)

Reasons Why I Chose These Articles

In an effort to ensure a fair comparison, I selected articles that had not undergone other significant changes over the last year, such as gaining backlinks.

I also ensured the articles were at least one year old to eliminate the URL age factor.

To explain further, the age of the URL is a ranking factor, as it can take months for Google’s algorithm to assign a rank to new content.

Although the changes in word count couldn’t be completely isolated due to domain-wide factors, like backlinks to other pages that strengthened the entire domain, including the articles in this study.

However, based on my 8-year experience in SEO, these factors were minimal, and the reduction in word count was the most significant change for these articles during the case study period.

Interpreting the Results and Advice for Site Owners

I was very happy with the results of this SEO case study.

They confirmed my assumptions that a more concise, yet equally comprehensive article would rank better than the bloated version of the same article.

Judging by the results, this effort was a smashing success for my site as there were many more winners than losers, and the wins were much more pronounced than the losses.

Of course, reducing the word count of articles isn’t always applicable.

It doesn’t make sense to reduce the word count of content that doesn’t suffer from bloat like mine did.

Each content publisher should make the judgement call to give this a shot based on their content.

One advice would be to analyze the competing articles from the top of the SERPs on a given topic and compare them with your article.

If your article sticks out like a sore thumb word count-wise, it might be bloated.

If in doubt, you can always perform an experiment on a smaller sample of articles and track the results before applying this to your entire website.