On March 29, I was a speaker at an SEO MeetUp organized by WordLift in Rome.
My talk was titled “By people, for people. Doing SEO in the time of AI. – A journey between E-E-A-T and GPT in search of quality”.
In this article, in addition to the slides from my talk, I would like to take a deeper look at the topics discussed to take stock of what role SEO plays today and what guidelines should be set for consulting work. I believe that it is important to take time to reflect, especially in a scenario that is changing radically and at great speed with the explosion of tools based on generative AI.
What is SEO?
When there are changes, it is important to go back to the basics on which our activities are based. Therefore, I would first like to establish a principle, which is also the vision that I try to convey to customers.
Usually it is said that SEO is a branch of web marketing, and one could certainly argue that this is indeed the case if we consider the usual situation where activities are aimed at sales and competition in the market.
However, I think this attribution falls short. For me, SEO has something to do with communication.
SEO is communication.
The typical scenario we have in fact is as follows:
- People have needs and express them through searches.
- The search engine is the tool that acts as a mediator. Today, Google, the most used search engine in the Western world, is more of an ecosystem than a search engine, because we no longer interact with the simple search bar, but there are different access points to a system of services through which our requests and searches are articulated – think Discover, News, Lens, YouTube, Maps, Assistant with voice search, etc.
- Websites serve a purpose and with their content and resources, they are meant to fulfill that purpose and satisfy the needs of users.
Among these communication topics there is SEO and there are us consultants.
Our work has to do with communication, because it deals with:
- Understanding how people search and what their intentions are based on the needs expressed.
- Understanding what the search environment looks like through its guidelines and thorough reverse engineering.
- Studying strategies and tactics for a site to best communicate its resources.
The goal is to meet the needs of both those who search and those who want to be found.
Regardless of the technological scenario in which we operate, this relationship is always valid. We have people with real needs who seek satisfactory answers and resources that can/will be found (and that we hope will be of quality or will become so).
I wanted to highlight these concepts because I have been in the SEO industry for a long time and have caught the hype around generative AI. It seems that the trend is to look for the easiest solution that can be effective in the short term, but at the expense of quality and satisfaction. This is a major weakness in an online project and means it will fail, if not harm it in the long run.
SEO is complexity and uncertainty, it’s a lot of ‘it depends’, but it’s certainly not a button you push that solves problems. So it’s not a tool to achieve default values, or a rule about how long an article has to be, or how many times a keyword has to appear and where. Google stopped thinking in terms of keywords at least a decade ago (see: introduction of the Knowledge Graph, things that are not strings) and yet, for many, SEO is still what it is.
I started working in SEO in 2010, and then as now, there is no one SEO recipe. And since we no longer have control over our resources in the Google ecosystem, it’s even harder to say: this is how it’s done. That’s why I want to share with you a method based on a vision that has been with me since the beginning of my work, starting with the concept of quality.
What Is Quality For Google
Although common sense can help us evaluate a resource, the concept of quality we have as humans still falls into the realm of opinions and subjective evaluations. What we are interested in with an SEO strategy is trying to determine how the concept of quality for humans matches up with Google’s criteria.
This is where Google’s official resources and, in particular, the Search Quality Raters guidelines help us.
We know that quality raters are human individuals who receive a set of Google results for evaluation. These ratings do not directly affect the ranking of websites or the SERPs, but they are used by Google to refine its algorithms.
The guidelines are a very detailed manual explaining the criteria used to rank pages and content. I find them very useful and interesting precisely because they offer a human-centric rather than a search engine-centric point of view.
The concepts we focus on when looking for quality are expressed by the acronym E-E-A-T.
Google E-E-A-T: What It Means For Quality
First of all, we would like to emphasize once again that E-E-A-T is neither a ranking factor nor an algorithmically measurable value.
For all SEOs who try to create always valid operating rules by reverse engineering: Sorry, but in this case it’s really not possible. (I have found some scripts in Python that promise with the GPT API to rank a content by E-E-A-T parameters, but as thought-provoking as they may be, I consider them real nonsense and we’ll see why in a moment).
E-E-A-T is a set of criteria where numerous factors come into play, the nuances of which require human judgment: There is no single type of quality content, but this changes depending on the purpose of the page and what it is intended to do.
According to Google, the quintessence of quality content is defined not by its type, but by its purpose: it must have a useful purpose, i.e. it must be useful for users.
This is even more important when it comes to content that falls into the YMYL (“your money your life”) category, i.e. content that has a direct impact on people’s lives, well-being, health and finances.
The work of a quality assessor consists in identifying the elements of the website, the home page and page types, its purpose and creator, and then determining whether the main content of a page under consideration meets the various quality criteria.
Let us look at them in detail.
E – Experience
Introduced with the October 2022 guideline’s update, it specifies the extent to which those who created the content have direct first-hand experience with what they are talking about.
Many types of sites serve their purpose well when created by people with rich personal experience. After all, who would trust a restaurant review if the person writing it had never tried the dishes or a product without testing it?
Indicates expertise in the subject, which means not only an academic degree or official recognition, but also informal expertise given, for example, by years of experience in the field. Think of a person who has been working in SEO for 15 years. She may not have a degree in SEO, but she can be just as competent if she has gained experience with numerous projects over time.
Expertise and experience can have overlap and nuance that can add value to the quality of the project, depending on the scope of the site. See the guidelines for examples, especially for YMYL content.
This factor concerns the extent to which the website or author is recognized as a reliable reference source on the subject. It goes without saying that while such recognition can be built through the development of expertise and experience, it also requires external recognition.
These three factors contribute to the construction of the most important value, trust: a website must be trustworthy and secure not only in the aggregate, but also beyond. To quote the example of Google: A financial scam is not trustworthy even if the creator of the content is an experienced and accomplished fraudster who is considered the best fraud expert!
When it comes to determining the quality of the site, the E-E-A-T rating refers to many elements and aspects:
- It takes into account what the site or the authors/authors say about themselves, in their profiles and on the “About” or “About Us” pages.
- It considers what others say about the site and the authors/authors through references, citations, independent news and reviews, articles, and other reliable evidence of the project’s experience, authority, and reliability.
- Assess what is visible on the site, from the main content to smaller areas such as the comments.
How E-E-A-T Is Applied To AI-Generated Content
With the advent of generative AI, many have begun to wonder how AI-generated content is evaluated by Google: is it indexed? Are they penalized? Can they rank well and thus be considered quality?
In February 2023, Google updated its guidelines with guidance on AI-generated content and tells us two important things:
– it does not matter how the content is created, what matters is that it is useful to people
– the automatic generation of content does not violate the guidelines if it is not used to manipulate the search engine and thus is not used for spamming.
Before chatGPT, there was already text spinning, i.e. the production of industrially produced and automated text, now this can simply become a more accessible and therefore mass practice. From this point of view, the discourse on quality and usefulness has not changed that much, even if the technological scenario around the available tools has already changed.
At this point, let’s see some practical tips and insights on what it means to create quality content that can meet the E-E-A-T criteria and what the roadmap of SEO activities could look like today, in 2023 and in the future.
Content SEO Best Practices
1. Read the Quality Raters guidelines
This document is definitely underrated, yet it is a valuable reservoir of guidance and advice that helps to look at things from the search engine’s perspective. Google does not always manage to deliver useful, high-quality results, but that does not mean it’s wrong to follow the guidelines provided.
In the given examples you will find many useful insights that you can apply to your own projects. The vision is always to work on the details, but also on the project as a whole.
In 2018, Danny Sullivan wrote a piece of advice on Twitter that I agree with:
2. Build authority through a solid content strategy
To become a reference source for your topic or area of interest, you need to be able to satisfy as many needs around your chosen topic as possible: a content strategy that covers as many topics as possible strengthens consideration for the source, because that’s where every question is answered. You are, so to speak, the “Wikipedia” of your industry.
Content strategies can be developed in many different ways and with different tools, depending on how we have built our workflow. This can be an interesting variable to consider as new technologies come on the scene, as we see with LLM and generative AI-based tools.
In the examples on the slides, we have an advanced tool, namely SEOzoom, which offers various tools for exploring topics, concepts, and keywords, such as keyword infinity, which allows us to cover a wide range of questions around a topic.
Then there are some AI-based tools such as Bing search in chat mode or ChatGPT itself for use with specialized prompts.
This strategy can really be applied in all contexts, from the most creative ones to those that are really difficult from a content point of view because they seem to be boring or of little interest, as it may be the case of a B2B company that deals with screws. But do not forget that people’s needs are diverse and can also have to do with… Screws to do!
3. Create relevance for search engines
After developing the content and resources by putting people at the center of the strategy, it is important that the search engine also understands what you are talking about. The information structure we have built and the project entities must remain relevant to the topics covered.
A good information architecture is definitely the first step to facilitate the exploration of content. This is true for humans as well as for Google crawlers searching websites.
Add to that the implementation of structured data markup, which gives the search engine a certain level of precision that it needs but may not recognize because it does not have the understanding that humans have. This task can be done by hand, but there are already advanced AI SEO tools that support and facilitate this task, such as the WordLift platform.
4. Building authority with human and experienced authors/writers
We can work on our value as content creators, but it is not always possible to cover all areas of knowledge. Therefore, it is important to enhance the project with content created by people who can demonstrate experience and expertise.
In this example, the healthline.com website displays a history of its editing and revision for each article, with names linked to author and author pages describing qualifications and experience. This structure is not only a sign of reliability and authority, but can also be supported by the specific structured data markup to increase relevance also from the search engines’ point of view.
5. Build authority with human relationships
I have long been of the opinion that link building activities can do more harm than good. You have to be very careful when choosing where to advertise, and often industrializing the process can have a negative impact on the outcome. Of course, this “depends” on a case by case basis. However, for Google, links and citations are of great value when it comes to determining the authority of a resource, just as the guidelines suggest we do: To determine whether a site or content is trustworthy, one must also consider what others are saying and how they are using the content.
Quality content, truly useful resources that solve problems in a practical and comprehensive way, is a rare value, which is why it can naturally become sources to cite and reference. It’s a reasonable process that may certainly seem tedious, but it can pay off in the long run.
Do not buy links, build human relationships and valuable resources to fulfill them.
6. Build confidence in the use of AI
We have seen that introducing generative AI into content creation processes can have benefits and effectiveness if the goal remains to create quality for users.
In the example, the bankrate.com website has chosen to create articles using automated techniques, to which it adds a process of review and editing by humans. Each step is transparently communicated to users, making a work-optimizing practice an added value: AI is used to make certain activities more efficient and allow a greater investment in quality, and communication becomes a trust factor for readers.
In a world where people are often deceived, this can become a differentiator and a signal of trustworthiness.
7. Ethics and responsibility in the use of AI
As consultants, we should always ask ourselves whether we are leading our clients to an easy but potentially dangerous solution for themselves and others, or to choices that preserve ethics and responsibility. There will always be those who operate with unfair or irresponsible practices, but we have the opportunity to choose and guide the decisions of others toward the values we believe are right for ourselves and others.
Google has a section dedicated to the principles of AI use, and I think the ethical debate needs to go hand in hand with the technical debate.
Aleksandr Tiulkanov, an expert on AI, data, and digital policy, shared on Linkedin a diagram of an algorithm for deciding if and when to use ChatGPT, and I find it really effective in its simplicity
8. Development of an algorithm-safe method
When we do SEO today, it has a lot in common with what was done before, when we worked from the perspective of quality.
So it is important to do the “same as before” and, if we have not already, to develop a solid method that can be applied in any scenario: In the era of AI, it is even more important to put people at the center, so as not to lose the communication principle on which the SEO activity is based.
So if we want to indicate an SEO roadmap, here is mine:
- Analysis of search personas and needs
- Analysis of topics/keywords and search intent
- Design of a robust and comprehensive information architecture
- Designing useful resources and content that match the search intent
- Writing content in a natural style (forget about SEO when writing)
- Creating micro-content: Metatag title, description, etc.
- Implementation of entities through structured data markup
- Technical improvements to all aspects of the site in terms of usability and user experience
Finally, investigating and developing new ways of working, such as integrating AI into repetitive and low-value processes that allow us to spend more time on quality.
This is my presentation:
Must Read Content
The Power of Product Knowledge Graph for E-commerce
Dive deep into the power of data for e-commerce
Why Do We Need Knowledge Graphs?
Learn what a knowledge graph brings to SEO with Teodora Petkova
Generative AI for SEO: An Overview
Use videos to increase traffic to your websites
SEO Automation in 2023
Improve the SEO of your website through Artificial Intelligence
Touch your SEO: Introducing Physical SEO
Connect a physical product to the ecosystem of data on the web