There were many expected causes of SEO’s death discussed in 2023.
One hypothesis that has been top of mind for me is the implication of Generative AI on the quality of traditional search engine results due to the quantity of content that websites will be able to produce. The evolution of tools like GPT-4 and Jasper AI empower users to create content at an unprecedented rate. This can be used both responsibly and irresponsibly.
- Responsible brands will integrate their unique data, insights and POV to cut down on time-to-publish, creating greater value for their audiences.
- Irresponsible brands will open the floodgates and exploit clever little schemes to drive views and impressions.
If you’ve been in the SEO industry for a while, you likely don’t find this inflection point to be shocking. SEO history is rife with both responsible brands putting in the work to win market share, and irresponsible marketers exploiting little cheats to get ahead.
That said, Generative AI is one of the most transformative technology developments of our lifetime. This may be the biggest threat Google has faced to ****. It’s fair to wonder: Are they ready for it?
As 2023 comes to a close and I prepare for 2024, I’ve been wondering what Google has up their sleeves to prevent this most recent inevitable downfall.
These reflections bring me back to the Panda and Penguin updates that guarded against content farms and link schemes, and the evolution of the engine during the Hummingbird and RankBrain updates. Every time individuals have schemed their way into unearned market share, Google has deciphered the scheme and eliminated the vulnerability. There’s been a lot of shade thrown towards Google in the past year about the degradation of their results. Rightfully so, but I’m still not ready to count them out. Surely they’ll come up with another new update, right?
After many hours of thinking and talking about this, I realized that I work closely with someone who has seen it all. Every update, every pivot, every notable advancement in the discipline of SEO through April 2023. I’m talking about GPT-4, of course. (Sorry, Wil!)
Before I asked for predictions, I wanted to make sure its knowledge base really was up to snuff. Here’s what GPT-4 gave me:
My Prompt:
Do you know about the history of SEO and how folks doing search marketing have tried to game the google algorithm over the years, and how google has responded? I am thinking things like webmasters investing in content from ‘content farms’ and then google responding with a focus on authorship. Could you give me this brief history from 2010 to present day, in bullet form?
GPT-4’s Response:
“The history of SEO (Search Engine Optimization) from 2010 to the present day is marked by a series of changes in strategies by webmasters and responses by Google to maintain the quality of search results. Here’s a brief timeline in bullet form:
2010 – 2012: Early Changes and Content Farms
- 2010: SEO during this period was heavily focused on keyword stuffing and backlinks. Webmasters often used these techniques to rank higher in search results.
- 2011: Google launches the Panda update. This was a significant algorithm change aimed at lowering the rank of low-quality sites, especially content farms (websites with high volumes of low-quality content). This change forced a shift towards higher-quality, original content.
- 2012: Introduction of the Penguin update. Google further refined its search algorithm to penalize websites that used manipulative link practices, such as buying backlinks or using link networks to artificially boost rankings.
2013 – 2015: Focus on Quality and User Experience
- 2013: The Hummingbird update was released, focusing on improving the interpretive capability of the search engine. It allowed Google to better understand the intent behind users’ queries, placing less emphasis on exact-match keywords.
- 2014: Google started to emphasize the importance of HTTPS, signaling a shift towards prioritizing security in SEO rankings.
- 2015: The “Mobilegeddon” update was introduced, which gave priority to mobile-friendly websites in search rankings. This was in response to the increasing use of mobile devices for web browsing.
2016 – 2018: Machine Learning and User Intent
- 2016: Introduction of RankBrain, a machine learning-based search algorithm component, which helped Google better understand the context of search queries.
- 2017: Google made numerous minor updates, frequently adjusting its algorithms to refine search result quality and user experience.
- 2018: The Medic update, which largely affected health and wellness sites, focusing on the expertise, authoritativeness, and trustworthiness of content.
2019 – 2021: BERT and Core Updates
- 2019: Google rolled out the BERT update (Bidirectional Encoder Representations from Transformers), which helped understand the nuances and context of words in search queries.
- 2020: Various core updates were released throughout the year, with a focus on content quality and relevance.
- 2021: Google announced the Page Experience update, incorporating metrics like loading performance, interactivity, and visual stability of pages (Core Web Vitals) into ranking criteria.
2022 – Present: Continued Refinement and AI Integration
- 2022 – Present: Google has continued to refine its algorithms with a series of core updates and improvements. There’s a growing emphasis on integrating AI and machine learning to improve search relevance and user experience.
- Ongoing: Google’s focus remains on high-quality, user-focused content, mobile optimization, page experience, and leveraging AI for better understanding user intent and query context.
This timeline highlights the evolving nature of SEO and Google’s ongoing efforts to improve the quality and relevance of search results, continually adapting to new technologies and user behaviors.”
As someone who lived through these changes, I’d say that’s a pretty good recap. So, let’s get into the predictions ChatGPT 4 made!
My Prompt:
“If webmasters start to use generative AI to create mass amounts of content, regardless of if that content is good and helpful or not, based on the history of google’s approach, how do you think google might respond?”
GPT-4’s Response:
Note that the format includes verbatim text by GPT-4 in bold, followed by my response in normal text. I’ll include a prediction grade for each idea as well, which grades both how likely this prediction is and how effective it could be in solving the problem.
1. Enhanced Content Quality Evaluation: Google could further refine its algorithm to detect and penalize content that appears to be generated by AI, focusing on uniqueness, depth, and context relevance. This could involve more sophisticated use of natural language processing (NLP) and machine learning to assess the value and originality of content, beyond just checking for plagiarism or keyword stuffing.
This would certainly be the most effective way to rid the index of AI generated content, but I don’t believe Google is willing nor able to implement this. First of all, I tend to believe Ethan Mollick when he purports that “AI writing is undetectable and likely to remain so, AI detectors have high false positive rates, and they should not be used as a result.”
Even if AI writing could be detected, Google isn’t inherently against the use of Generative AI to create content. When used safely, effectively, and thoughtfully, (ahem, see: the P-A-R-S method) Generative AI can be a powerful tool. I don’t believe Google will go full bore against Gen AI content anytime in the near future.
Prediction Grade: D
2. Emphasis on Authorship and Credibility: Google might place greater emphasis on authorship and the credibility of content creators. This would mean prioritizing content created by verified human authors or reputable sources. Google could develop more advanced methods to verify the authenticity of authors and sources, potentially integrating these metrics more deeply into its ranking algorithms.
The history of Google Authorship is well documented in this post by Eric Enge on Search Engine Land. While the original iteration of Authorship is long dead, the concept is back on the rise.
This aligns beautifully with Google’s consistency around E-E-A-T principles. As we’ve dug more deeply on E-E-A-T, the need for authorship verifiability has become greater. Google could weigh these elements more heavily through tactics like Author Bio pages, Author Bylines, Author Schema, and more.
Prediction Grade: A-
3. User Engagement and Interaction Metrics: The search giant could increasingly rely on user engagement metrics such as time spent on page, bounce rates, and interaction rates to gauge content quality. Content that engages users meaningfully would be favored, whereas content that fails to retain user interest (regardless of being AI-generated or not) could be ranked lower. This approach would naturally filter out low-engagement, AI-generated content that doesn’t meet user needs or interest.
This is a great theory, and one that builds upon a core element of Google’s algorithm today. Thanks to the documents revealed in Google’s antitrust lawsuit, we know Google factors in post-click metrics as part of their ranking process.
This prediction would have Google pressing it’s thumb on a pre-existing scale, making it easy to implement quickly. The downside is the same as the downside that exists today: This ranking factor tends to enable the rich to get richer, since higher ranking sites have more post-click data fed into the index.
Prediction Grade: B-
4. Encouraging Rich Media and Interactive Elements: To counter the flood of text-based AI content, Google might start favoring content that includes rich media (like original images, videos, and interactive elements). This shift would encourage webmasters to focus on diverse content forms that are harder to replicate with AI, thereby enhancing the overall quality and user experience of web pages.
From a usability perspective, generally this sounds like a win. One implication of the volume of content being fed to us every day is users are beginning to self identify how they prefer to ingest content, and publishers are accommodating. Sometimes I want to dig into a long read, other times I want to hear the audio version, and the need for visuals or video is highly dependent on the topic in question.
That said – this would be a band aid solution at best, as the text-to-anything space is growing at a very fast clip. Today, creating custom visuals, audio, and video at a fast pace is possible. In 2024, we’ll be blown away by the advancements.
Prediction Grade: C
5. Rewarding Verified Expertise and In-depth Research: Google could place a higher value on content backed by verified expertise or in-depth research, such as academic citations, expert quotes, and comprehensive analysis. This approach would reward content that demonstrates a deeper level of knowledge and understanding, which is typically beyond the current capabilities of AI content generators.
Everything old is new again! I was pleased to see GPT-4 reached the same conclusion that some SEOs are predicting themselves. With a flood of content hitting SERPs, will Google adjust their algorithm to weigh links more heavily?
In this case, we aren’t talking about external backlinks. Instead, this mechanism would give credit to those who are finding a way to integrate citations into their content. It’s a good theory, but based on what I have already seen with Generative AI, I think it would be very easy to insert these references and citations at scale.
Prediction Grade: C-
6. Focusing on Contextual and Behavioral Signals: Google might enhance its reliance on contextual and behavioral signals to evaluate content quality. This could include factors like the context in which the content is shared, user interactions (comments, shares), and the content’s performance in related searches. Such signals can offer insights into the content’s real-world relevance and value, which might be less effectively replicated by AI.
Back to the future, once again. One of the main downfalls of links as a democratic sign of support is that the average individual doesn’t have the ability to ‘vote’ unless they happen to control a website. Google spent a lot of time and effort trying to factor social signals into their algorithm, but unfortunately most social platforms have the opposite problem that backlinks have: It’s far too easy to manipulate the vote.
This hasn’t been a major talking point in years, and Google has only gotten stronger here. I would not be shocked if verified social proof was connected in some way, shape, or form to authorship and rankings in general in 2024.
Prediction Grade: B
7. Dynamic Content Analysis: The search engine could shift towards more dynamic, real-time analysis of content. Instead of relying solely on traditional metrics, Google’s algorithms could assess how content evolves over time, how it is updated in response to new information or user feedback, and how it interacts within a broader content ecosystem. This dynamic approach would favor content that remains relevant, updated, and interactively engaged with its audience, characteristics that are challenging for static, AI-generated content to maintain.
This is a fantastically interesting concept. The act of creating tons of content at scale with the right mechanism is unfortunately pretty easy. The act of updating tons of content at scale? I’m not sure how easy that is. This concept aligns well with Google’s push to provide more real-time information to users, and would help to improve user experience with SERPs overall.
Prediction Grade: A+
In Conclusion
This post helped highlight a few things for me, and I hope it did for you as well.
- ChatGPT is a great thought partner, and I am excited to use it more for hypothesizing and scenario planning
- Google has many avenues to take in order to mitigate risk to the integrity of their results, and I would feel confident ******* that they are already pursuing a few of the ideas ChatGPT laid out above
- All that is old is new again, and that includes how we believe marketers should prepare for major algorithmic changes.
If you are wondering where to go from here, I would give the same advice we have been giving for over a decade: Keep creating content that your audience finds value engaging with. Keep making your website accessible to search engines. Keep ensuring that your site is well optimized for the topics your audience is searching for. The only new addition I would add is to consider how Generative AI can help you do all of the above, and then some. Think of the tool as a copilot, not an autonomous machine, and you’ll be able to maximize its impact while minimizing its risk.
