Hi friend! 👋
We’ve all heard the advice: write for people, not robots (search engines.) It’s sound advice, and if you noticed Google’s latest algorithm update, there’s no question about it.
But how does the logic apply if a robot does the writing? Should a website be penalized for making use of (pretty amazing) technology?
In today’s edition of Advance Insider, we’re going to dive into:
At the end, we’ll wrap it up by sharing our current open positions! 😍
What does helpful content mean?
To us, helpful content is content that helps (hehe, funny that) your target audience achieve something.
Perhaps that’s learning how to do something, learning about something, or solving a problem they have in their business. We know that content (whether it be video, visual or text-based) is the best way to reach and help target audiences for most B2B SaaS and subscription-based companies.
But how do we know it’s helpful?
Well, by making conclusions based on data, of course. If a piece of content is successful (say, in terms of high-quality traffic, leads, free trial sign-ups or demo requests) we can also assume it helped somebody get closer to solving a problem, and thus sending them further down the funnel — and closer to your product.
About SEO content more specifically, though, there still seems to be a lingering opinion that content for discovery on search engines doesn’t always need to be high quality. Or that writing content for search is somehow different than writing content for your ideal target audience.
This is interesting, because it’s clear that Google will penalize (or at least not reward) a website for not producing people-first, valuable, purposeful content — they said it pretty clearly in the Helpful Content Update release notes.
Some ‘red flags’ from the update that should prompt websites to review their content include:
So what happens now?
As this article by Search Engine Land points out, it seems there hasn’t been a huge change in SEO performance for the majority of websites. Which is a great thing! The effects of the Helpful Content Update may be amplified with the impending September Core Update, but as long as you didn’t answer ‘Yes’ to the questions above – you should be all good.
It is a good reminder, however, to consider reviewing the purpose behind each piece of content we choose to create – and consider exactly how it satisfies our audience, or helps them get a job done.
This could be in the form of a retrospective content audit (which we, by the way, just launched as a standalone project) or as a part of your standard content creation process going forward.
Here comes the but…
But in saying all of that…what if we could get a little help from AI?
Enter: GPT-3 (the newer, actually pretty decent version)
GPT-3, or the third generation Generative Pre-trained Transformer, is a neural network machine learning model, developed by OpenAI.
The model is trained using internet data to produce, autocomplete, summarize, or classify any type of text. It only requires a sentence or two of prompting to create large volumes of text – and the crazy thing is, it actually makes sense. Most of the time. Previous versions of GPT were shockingly average and didn’t provide much value, but the latest update (formerly known as InstructGPT) is something remarkable.
Take a look at this example:
Pretty neat, huh? And that’s only a tiny snippet of what it’s capable of.
Now, the things GPT-3 can do are pretty incredible. But we must approach AI-based writing tools with caution, particularly when using them for content creation.
🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮
For our Finnish-speaking audience, Sari wrote about some tools that work for Finnish speakers last year – find it here! 🙌
🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮 🇫🇮
An important thing to consider is what GPT-3 can’t do (yet?):
See what we’re getting at here?
The challenge lies above all in the fact that artificial intelligence is ultimately always based on some existing material, and thus it is not possible to bring the kind of insight and vision to the article that requires proper thinking. Just a credible-looking wall of text is not enough — human brains and critical thinking are still needed for strategic content.
GPT-3 is a powerful tool for generating small snippets of content for very specific use cases. With great power comes great responsibility, and in the case of creating helpful content (which should be our ultimate goal) it’s not quite there yet.
Know a Hubspot CMS Developer?
If you, or someone you know, is a Hubspot CMS Developer dreaming of an inclusive, learning-obsessed, fair and flexible workplace (full of awesome people) you’ve hit the jackpot: check out the position here.
We also have some other open positions for non-developer folks too.
Could we be the right fit for each other? These should give you a rough idea:
And about that podcast episode…
To hear about GPT-3 and content from somebody who has spent a great deal of time learning, testing and analyzing the technology from a content marketing perspective, tune in to next week’s episode of The Growth Hub Podcast.
Our hosts, Seija and Reeta, chat with Ryan Law, VP of Content at Animalz and staunch proponent of GPT-3.
Subscribe to the podcast on Spotify, Apple Podcasts, YouTube (or wherever else you get your podcasts) to be notified when the episode drops.
In the meantime, we’d like to hear from you!
What are your thoughts about AI-based writing tools like GPT-3?