ChatGPT vs. Marketer: Who Crafts the Superior Thought Leadership Article On LinkedIn

Can ChatGPT replace you (or me) as a writer, expert, and thought leader?

That’s the question on everyone’s mind.

While AI writing tools aren’t brand new, their popularity has grown in the last year.

According to Y Combinator Jasper, an AI writing company, was created back in February 2021. But, it wasn’t until OpenAI released ChatGPT at the end of 2022 that things started to change.

ChatGPT reached 100 Million monthly active users in January 2023. Contrast this to Netflix which took 3.5 years to reach 1 Million users.

Over the last year, AI has taken over our social media feeds as we’ve watched 100s of AI companies get launched, people capitalizing on the hype, and rapid advancements in the AI arms race.

This has quickly led to the question surrounding jobs…

…Who’s safe, who’s not, and whose jobs are most at risk?

Since ChatGPT can write content so quickly, many have begun to lean on it to write social media posts, SEO blogs, and thought leadership articles.

But, are they better than if we wrote them ourselves?

So, I decided to test this out for myself.

Every Saturday morning I sit down at my laptop and spend 2 hours writing.

This is my “long-form” writing container.

Going on 6-months I’ve been writing thought leadership articles on Substack and LinkedIn sharing my experience and knowledge.

I’m no stranger to ChatGPT.

Back in 2022, I created a TikTok on writing headlines with ChatGPT that got over 230,000 views and I’ve published several articles, videos, and workshops on using it as a ‘helper’.

But what I’ve never done..?

Is have ChatGPT write an entire article for me.

It was time I tested ChatGPT to see who could write a better (and more engaging) thought leadership article for LinkedIn.

Now, I’ll walk you through my entire process.

From the prompts that I used, to where the articles were posted, and finally the results, which might surprise you…

If you’d like to read each article, you can find them here:

My article: 7 Expert Tips To Implement Into Your Facebook Ad Strategy In 2024

And, the post that shared this article can be seen here.

ChatGPT’s article: Mastering Facebook Ad Marketing: Strategies for Success

And, the post that shared this article can be seen here.

The Outline Of The Experiment And The ChatGPT Prompt I Used.

The goal was to create a thought leadership article on Facebook Ad strategies.

I replicated my typical process:

When I sit down to write, I start by prepping the page.

I map the rough idea of the headline, followed by the main points/subheads, and this gives me the structure for what I’m about to write.

Once I completed this, I passed it along to ChatGPT.

My prompt included:

• The goal of the article.
• The length of the article.
• The structure of the article.
• The main points of the article.
• And, how I wanted to break down the information for the main points.

Here’s a screenshot of my exact prompt:

After generating the article I saved it and proceeded to write my version.

I didn’t read it or analyze it first, because I didn’t want to skew my thoughts around what I wanted to create.

If you’d like to read both articles you can find them here:

• My article: 7 Expert Tips To Implement Into Your Facebook Ad Strategy In 2024
• ChatGPT’s article: Mastering Facebook Ad Marketing: Strategies for Success

Here’s How I “Designed” The Experiment.

To keep the results as “fair” as possible I decided that I would post both articles across social media using the full text.

I did this to keep people on-platform rather than posts that linked back to the original article. That throttles reach.

(I will caveat this by saying I don’t have a large social media following so the results are relative to my respective reach)

Each article was posted on:

• LinkedIn, as an Article.
• Twitter, full length since I have a premium subscription.
• And, Facebook, linked back to LinkedIn which was a mistake.

(Note my Facebook reach was non-existent so it’s not included below)

My version was also posted to Substack but I didn’t want to spam my subscribers with 2 versions of the same article so I didn’t included any Substack views in the results.

Each version had an image associated with it:

For my article, I used my typical style and prompts to generate an image that resembles my normal theme. And for ChatGPT’s version, I just pasted the text it generated and asked it to create an image for it.

My Image:

ChatGPT’s Image:

Each version had a headline associated with it:

I wrote mine before I generated the ChatGPT version so that I was not influenced by it’s version. And ChatGPT generated the headline for it’s article.

My Headline: 7 Expert Tips To Implement Into Your Facebook Ad Strategy In 2024

ChatGPT’s Headline: Mastering Facebook Ad Marketing: Strategies for Success

Alright, with the frame set…

Here Are The Results…

My Version

7 Expert Tips To Implement Into Your Facebook Ad Strategy In 2024

LinkedIn:

• 200 impressions
• 1 reaction
• 2 comments
• 13 article views

Twitter: https://twitter.com/LandonPoburan/status/1759623339835031879

• 39 Views

ChatGPT Version

Mastering Facebook Ad Marketing: Strategies for Success

LinkedIn:

• 381 impressions
• 14 reactions
• 16 comments
• 17 article views

Twitter: https://twitter.com/LandonPoburan/status/1759977999896867160

• 20 Views

One thing that the impressions and engagement don’t tell you, though?

How long it took to create…

My article took me 2 hours to write.

On the other hand, it took me 10 minutes to generate the ChatGPT Article.

Whelp… I Lost. But Here Are My Thoughts About The Experience.

I know, know.

My results are not statistically significant.

While I would love to have 100,000 followers and be able to pull some big numbers to compare…

I gotta’ work with what I’ve got.

You’re welcome to follow me to help me work towards this ;-)

In short, this was a fun experiment, and here are my 4 takeaways.

Takeaway #1: ChatGPT was significantly faster, and had a similar (better) “Reach”.

2 hours vs 10 minutes is a big deal.

And it technically got more reach. Not enough reach to know if this would translate to real-world sales or audience growth, but the similarity is worth noting and further testing.

Takeaway #2: The content was 70% of the way there.

The prompt matters.

Providing the theme, length, structure, and primary points got it most of the way there. The overall theme was captured and the main takeaways were in line with what I wanted to get across.

Takeaway #3: It didn’t capture my voice or the way I’d personally write it.

It was a bit hypey and “off-brand”.

With better prompting, I’m sure the style could be improved. But ChatGPT still writes things in its own way. As ChatGPT evolves, or through the use of Custom GPTs trained on my content, I am sure this can evolve or be improved.

The question then becomes, how much of our “voice” matters for personal branding.

Takeaway #4: It focused on facts.

My article included practical examples whereas ChatGPT was straight facts.

If I provided an example or asked ChatGPT to include a story or an analogy with each point, this could change. But the one thing AI cannot replicate (yet) is my own personal experience, and my clients’ results.

Lastly,

I do think the ChatGPT article had a headline. It led with “Mastering Facebook Ad Marketing” which I think was stronger than what I wrote.

Here’s What I Think Is Most Interesting From This Test.

What if I could train a Custom GPT on my writing, my voice, and my client results…

Uploading all of my articles, podcasts, YouTube videos, and social media posts.

How close could it be to capturing my voice - and the stories and results I reference?

I know some people are doing this, and I hope they do a similar test.

While I won’t be replacing myself completely just yet, I will continue to use and test ChatGPT and AI.

I consistently use it (and will continue to) for information, ideation, planning, headlines, images, and more.

But for now,

I will continue to write my own thought leadership articles and use AI for support.

Even though I lost this one…

I still love the process of writing and processing my thoughts through the written word.

For me, that means more than the data. For now ;-)

That’s all for now, until the rematch.

-Landon