The future of SEO copywriting

The latest post by Andrew Ng on LinkedIn goes like this:

ChatGPT is sometimes amazing, and sometimes hilariously wrong. Here, it explains to me why an abacus is faster than a GPU. 😀

It is hilarious how AI can reason about topics faulty from the beginning.

Let’s rewrite the last passage about abacus and CPUs in a simpler manner: abacus is smaller than your server rack, you can even put it in your pocket, therefore it’s better for deep learning. Doesn’t make sense, right? But when written with more obscure wording, it may sound persuasive for people who don’t know much about computers.

The NG’s post made me think about loads of low-quality texts online and whether the situation is going to get better anytime soon. Here I listed my thoughts on the issue.

Most copy these days isn’t any good

What you can find online about bad copy just isn’t true. Most articles say that the traits making your writing bad are typos, plagiarism, not understanding the target audience, and jokes that went wrong. Of course, these are bad, but who cares about such errors if your text is actually interesting to read?

Deriving from my experience, the most widespread traits that make texts bad are:

• factual errors
• outdated information
• obscurity and verbiage.

Why average copy is so bad?

Now, blog writing is outsourced to countries like India, where an average writer is forced to produce an insane amount of text per day to pay for their needs. Averagely, she covers a vast range of topics in order not to be short of orders. The number of domains the author writes about is too large for a person to be an expert in each field. That’s how we get low-quality content online, tons of it.

But it’s not only about the pay. Often, clients come to writers and ask to create a piece on a topic that makes as much sense as comparing apples to tablespoons. This is because topics are proposed by marketing strategists and are aimed at featuring the needed keywords.

This leads me to discuss another pitfall: technical staff doesn’t check blogs for bullshit. It’s only the content manager’s and editor’s responsibility to go through the texts. But these people often don’t understand the topic well enough to spot factual mistakes. This is how something like “abacus is better than a GPU” can make its way to publication.

Then, often clients ask to frame an outdated technology as a modern one, as the company is offering services involving this technology.

That’s not it: often, there’s not enough input from the tech staff. They probably believe that blogs are created by some extraterrestrial force that has an omnivision ability to read their minds.

The future of copywriting

So what are we to do with the situation? How things are going to evolve? In my opinion, there are two scenarios:

• Either copywriters will die out because GPT-3 is good enough, and the Internet gets even more flooded with generated texts that make no sense, like this one explaining why an abacus is better than a GPU unit for deep learning, or
• Writers will obtain expertise in a number of specific domains and, therefore, will be able to write better texts than GPT-3, spot factual mistakes and correct them.

However, for the last scenario to come true, writers need to be offered a better pay. But I believe that businesses have no interest in publishing better blogs. As long as low-quality blogs can get ranked high by search engines, companies will continue publishing them. It’s cheap and effective, so why not?

Way out

Let’s hope that Google will start ranking pages using more sophisticated metrics, and in doing so, it will nudge businesses to hire experts to write better blogs.

Possibly, new metrics will also lead to the demise or significant shrinking of the marketing strategy involving SEO-oriented blogs and landing pages.