Content Marketing & ChatGPT

February 13, 2023 by Michael Spinka

Recently, I’ve had several discussions with colleagues and clients about what ChatGPT and A.I. copywriting means for their content marketing strategies, and what is to come in the future now that A.I. can produce long-form marketing articles. There’s a perception among those that pay for content writers that the A.I. will simply do it for them. This is my response in the short term as the marketing universe begins to adapt this technology into more widespread practice.

A content writer produces marketing copy using A.I. tools like ChatGPT.

Chat GPT is exciting; it’s the new thing. It’s built on the Open A.I. engine, and my agency has been using A.I. generated tools for a year now, born from the same engine. So far, my experience is that they’re great at opening lines and first paragraphs, but have shown that the deeper you try to go on a topic, the A.I. starts to spin the same openers, ending up with mostly unusable long form copy. I equate it to a speaker that started off with a strong thesis, and then spent a half hour speaking in circles, never providing any real depth or proof into the topic. The A.I. doesn’t know where to go next because it’s not an expert. That will continue to evolve though. Since the real depth it lacks comes from practical authority and experience, the question is how much (and how fast) can the A.I. learn about expertise while not actually having any?

According to Google, A.I. generated content is against their guidelines, even though they have their own chat (Google Bard A.I.) that was just released. For years they have been dealing with content marketers posting borderline gibberish articles on the web that tease in a few keywords here and there attempting to get the content to rank in search results. This over-publishing of keyword laden content pre-dates A.I., where marketers would take articles, and place them on hundreds of websites in an effort to trick Google into crediting sites that had backlinks in the content, so those linked pages would rank a little higher. They call it passing “link juice”. But you can’t place that same article on hundreds of sites. Google knows when it sees duplicate content posted. So they would spin some of the words replacing adjectives and such to throw off the index scanners to think each instance of this article was unique. Slick, huh? Another hurdle for Google to catch among the many illegitimate practices of trying to get pages to rank in search results.

Recently Google released a “helpful content update“. The idea is to promote sites that produce helpful content to the readers, and de-index, or demote the spun text junk content. How will they know if content is helpful? One easy indicator that comes to mind is they already know when people bounce off a site and go back to the search results and click another site from the result set. That is just one example that says the content the user clicked first may not have been helpful. I’m sure they can determine fluctuations in normal people bouncing due to researching multiple outlets of information vs bouncing quickly from poor content. I don’t claim to know how their algorithm manages it, but it’s commonly accepted that it’s not the first click from a search engine result that’s the most important…but the attribution from the last click. The point here is that even with the development of A.I. content, it has to be helpful, and contain knowledge and expertise or it will just get picked up in the filters that bypass the junk when it comes to ranking. For a lot of marketers, that’s why we produce so much article content. It’s valuable toward incremental ranking opportunities that are part of a larger SEO growth picture.

So what actually is their “helpful content” guideline? A solution called E-A-T. Expertise, Authoritativeness, and Trustworthiness. Content marketing should be built around these elements. The A.I. does not appear to be able to beat this…yet. Rest assured when it does, the accessibility to push out A.I. generated content will be tremendous. Abundance by the second. Articles, articles, and more articles. As marketers, our goal is to get attention and keep it, which allows us to nurture our leads into a pipeline. It will be harder to get that attention when the marketplace is filled with endless content spun by computers and presented as experience and authority on a topic. We are trying to win the trust of our prospects. Trust does not happen at the click of a button. It’s just not that simple.

What we know now is that ChatGPT and other prompt based A.I. tools should be used wisely, as assistants to humans that can add to it, a level of E-A-T that gets through to their buyers. A.I. tools are presently great for taking a blank slate and helping to start an engaging article, but the legwork requires real knowledge and experience to use and leverage the information effectively.

A great example of what is likely to come next, we can learn from B2B social media memes. 3 years ago, everyone was posting memes on their social channels. Stats, quotes, facts, Q&A, etc. Then the content aggregators started mass producing graphics of memes that received engagement from other marketers. People start copying and duplicating the same stuff over and over. Then they sold those posts in bulk – 10,000 posts for $50…WOW! Now our social media is all set for the next 10 years! Woo Hoo!!! Where do I sign up? And like a flash in the pan people stopped paying attention, and those posts for the most part went away, because the reality is that social media is more about people than posts. The same applies for customer acquisition being more about trust than the amount of content that appears on a website. And many of the sites we manage, the “About Us” is one of the top viewed pages because buyers want to know who they’re learning about just as much as what they may need from that organization.

In social media, when there is abundance of low engaging posts, it just clogs the feed and then followers ultimately unfollow. The same applies for other content marketing especially the kind that we use for SEO to capture website visits from those searching for topics to which we are experts. Those individuals are seeking knowledge on a topic, and they expect the search engine result to provide adequate information that answers their query. If users of google just keep landing on computer generated content that is NOT useful, two things happen. One, the searching user learns the site they visited was not useful and they are less likely to return again. Two, it says Google did not provide an adequate result, and that the effort to find information was not successful. That’s the level of expectation we all have from Google, and if they can’t deliver then people will stop using them. This is a genuine concern for Google as Gen Z prefers social media and video to search and learn about products over written content by marketers. They view it as being more authentic than reading articles. An entire generation already prefers methods of authenticity across channels the A.I. can’t effectively infiltrate any time soon.

This new technology is pushing toward an evolution; not a solution. I’m excited about it too, but will encourage clients and associates alike to continue to focus on what gets through to buyers, and Google already figured it out…E-A-T. While some other marketers are worried about what this will mean for content writer jobs, the future for now seems that the A.I. has a long way to go, but A.I. is proving to be a valuable tool to use along the way.

Happy publishing!

Michael Spinka has nearly 2 decades of demonstrated success in marketing & technology, and was formerly a vice president with a #1 marketing franchise that scaled coast to coast.

He founded The On Fire Group to help small and medium sized businesses reach their target audience affordably. Connect with Mike on LinkedIn. Subscribe to stay up to date!