I was in Turin for WordCamp Europe last week, the largest annual WordPress conference.
I found it noteworthy that attendees’ approaches to SEO varied depending on their sites’ performance. People facing traffic declines voiced futility and hopelessness, whereas those seeing positive trends (which naturally included many of Ellipsis’ partners) expressed much more optimism and confidence in the future of SEO.
As we’ve discussed in recent newsletters, SEO is changing and I relish the challenge. This is a period of disruption and we’re perfectly placed to capitalize. The changes to search value quality, originality, and insight. Previously you’ve been able to get away without those.
Hence: the tide is going out. Who has pants on?
Even good content may not be enough to get search performance
As you’ll know as a reader of Searching for Answers, Google search is changing. We know the headline themes: focus on quality, expertise, and insight. We also know that a huge amount of content does not offer these and will lose out.
Google always has a large presence at WCEU. Noteworthy for us was a rare appearance by Danny Sullivan, Google’s Search Liason, with whom James and I were grateful to chat. My main takeaway from the conversation was Danny’s take on “good enough” content.
For a couple of months, Google has simultaneously demanded content quality while rewarding content of inconsistent quality (especially for generic searches, where Reddit threads are more prominent).
This has led to the rise of “Goog enough” content from some larger brands such as Forbes, where the content is pretty bad but it clears a low bar for quality. This option seems to be only available to bigger brands.
In the long term, this content will surely fall away in favour of higher-quality competition. AI improves Google’s ability to understand content and with this the ability to better reward more relevant and higher-quality content.
WCEU was in Turin and I, like many attendees, Googled “what to do in Turin”. Danny had done the same. Much of the content was perfectly good, but it also struggled to bring any novelty to the table. Given this, good and insightful content may not be sufficient to get a ranking as there is plenty of other good and insightful content for Google to consider.
If your post is great – but there are also many other great posts – then being great may not be sufficient.
Topic coverage and the complete guide-ification of everything
Improvements in Natural Language Processing (NLP), a type of AI, in the late 2010s allowed for improved programmatic understanding of what’s in a piece of text.
This enabled Google to better understand webpage content, crucially equipping it to better serve content to the right searchers. A couple of entrepreneurs realized the opportunity: they, too, could use NLP to understand the content of web pages.
The theory is simple: aggregate the results from the top 20 pieces of competing content for a target keyword, and you can easily access a list of the specific topics covered. If you’re trying to create a more comprehensive article than your competitors, this lets you judge what you need to do to make this happen.
This “complete guide-ification” of everything was a very good idea for a couple of years, but it quickly breaks down in a genAI world. Pre-ChatGPT, the cost of creating a long article was sufficiently high that you could use article length as a proxy for quality. This is no longer the case.
Suddenly, the idea that you’d create a very long article cheaply and quickly is possible. If you’ve got a list of topics to include, even better. The trouble is if you extract topics – as this software does – at the word or short phrase level, you lose all meaning.
An article about coffee grinders might return “You should talk about ‘coffee beans'”. You have no idea what the searcher actually wants to know about the topic. If you throw genAI at the problem, you can generate a lot of content that includes the right words but in the wrong order. This is not so useful.
It’s time to move away from what everyone else is doing
My takeaway from speaking with Danny reinforces the direction of travel we’ve seen so far this year: it really is time to move away from what everyone else is doing.
Identify from first principles what the searcher wants to know and what unique angles you can bring to a topic, and that will likely give you a different result from what everyone else is doing. And that’s better.
This makes content production and SEO harder as it requires actual expertise and insight. We’re working on a new AI research tool that fits this philosophy and we hope to share more on this soon. We’re going to need new innovations to make this kind of content easier to create at scale.
It’s time to adapt.
