The SEO life is not for someone who desires to work quietly from 9 to 5.
And for the past 3 months, update after update and SERP’s test after SERP’s test (and now with the roll-out of Gemini, Google reminded us of this.
With such a volume of changes, it is natural that many SEOs openly declared their despair. However, desperation – albeit understandable – is not a solution, and the best reaction is to breathe and try to look at Search from a higher perspective and rediscover the value of strategy.
If we think in strategic terms, and if we apply Occam’s Razor, the ultimate purpose of any SEO strategy is to generate the biggest volume of qualified organic traffic toward the correct pages of a website to result in conversions/leads.
However, if until not so long time ago understanding the search journey of our potential customers, and so guiding them along it, was relatively easy and predictable, nowadays it is not so anymore, because they can follow a substantially infinite number of paths before landing to a web page (note: I say web page, not website…).
Aside from all the discussions about the “battle of dominating the AI-space” and the changes in how people search for something on the internet, I believe that the frenzy we are experiencing on the SERPs is a classic example of “Gattopardismo”: to change everything so that everything remains as it was before.
SGE, for instance, is nothing but another step from Google in the direction of offering all the answers in the Messy Middle inside its ecosystem.
If you have been doing SEO at least for a few years already, you know how all of this has happened before, and know that all of this will happen again.
But, citing the meme, “one does not simply enter into Mordor” or, at least, not so easily as before.
Nevertheless, what if Google itself could help us navigate the uncertainty, and help us design a strategy that can make our websites almost immune from all the changes?
I think this is a real possibility, and I will try to explain why in this post presenting you an exercise for a niche I know very well: mini-soldier tabletop games.
Please, note that even if the “virtual website” I will talk about in this post is an ecommerce, practically everything I am going to present here is valid for any kind of website.
Search Journeys
As said before, users now can travel along an infinite number of search journeys in the Messy Middle, substantially breaking the old classic pattern of the conversion funnel.
However, it is not realistic to try to discover and target all the possible search journeys. The correct strategy is to individuate the most probable ones and once individuated, review the foundations of our website and web presence to be always outstandingly visible along the biggest number of these paths.
The first secret hidden in plain sight: the search menu.
Us SEOs usually do not pay attention to the Search Menu because it is just that thing that allows us to navigate from one vertical search to the other.
Very few know why Google changes the order of the links in the menu from SERP to SERP (“It’s another Google test” is the classic answer).
The answer is: that Google may change the order of the filters based on the knowledge it has about people’s behaviour when using a given query set.
Let’s see an example:
If someone starts his search journey with a branded query like “Star Wars: Legion”, Google statistically knows that his intention most probably is to buy that tabletop game (“Shopping”) and that he tends to prefer visual shopping to other options (“Images”) and, eventually, with the help of videos reviews of the game for taking a final shopping decision (“Videos”).
The value of the classic search filter order analysis is clear, especially for understanding the real primary Search Intent and the ideal content formats to use to satisfy it.
But… these are not the only filters Google presents to us in the Search Menu. Another type of filter is represented by the Topics.
What are the Google Topic filters?
As explained by Google in its official documentation, Topics allow us to add terms to our query that can help us get more specific information or explore related information.
Relevant topics for a query are automatically generated and displayed, and they are based on what the systems of Google understand about how people search and how content is analyzed across the web.
In plain English, Topics suggest to us which are the statistically more probable search journeys someone can take from the starting point represented by a query.
In our exercise, the Topic filters for the query “Star Wars: Legion” already offer us great insights.
In fact, along with obvious potential search journeys (“Star Wars: Legion new releases”, “Star Wars: Legion cheap” or “Star Wars: Legion rules), Google presents us with the topic “Discontinued”.
Discontinued products are very important in tabletop games with minis because these games collide endlessly with the collectable niche due to the passionate fanbases they can create (think of Warhammer, for instance). Knowing this potential search journey, we can immediately recommend not eliminating old product pages and redirecting them to new products or categories, but creating an “discontinued” catalogue section or faceted navigation and, eventually, offering customers and fans to upvote the discontinued units they would love to see returning in production.
Therefore, a simple look at the Topic filters already highlights a facet that we should not overlook when creating an SEO strategy for a website (official or affiliate) about “Star Wars: Legion”.
Then, Topic filters can be analyzed as we analyze the questions of the People Also Ask search feature.
If we click on a Topic, we are sent to a new SERP, which presents new topics:
Each Topic, as said, corresponds to a new query. These are the ones we must note down in the sheet we are creating for noting down all the insights we are discovering while analyzing the SERPs:
A good idea for discovering even more detailed and nuanced potential search journeys is to scrape the Google Suggest for each Topic’s query.
For instance, the Google Suggest for the “Star Wars: Legion Clone Wars” show us another facet that usually is undervalued by e-commerce/websites of tabletop games: tools helping players build armies.
The second secret hidden in plain sight: related searches.
Related searches are nothing new but, maybe because they are presented at the bottom of the SERPs, they tend to be completely ignored by SEOs.
Although, they are a great source for understanding the potential Search Journeys.
Why?
Because, as it was for the Topics, Related Searches are automatically generated based on what our systems understand about how people search.
As we can see here above, a relatively easy analysis of the related search chain for “Star Wars: Legion” shows us how people tend to go on Reddit for news about new releases for the game, and the specific new releases they are most interested about hence the ones we should prioritize in our internal linking optimization and also with creating content able to target the informational content strongly based on experience with those new units too.
The third secret hidden in plain sight: SGE “Follow up” questions.
SGE, which now is starting to be fueled by Gemini, is (going) to be another important starting point for a Search Journey inside the Google ecosystem.
One of its features is the Follow up” questions:
When we click on a Follow Up, the SGE will expand the conversation with the AI, and this will present other more detailed Follow Up questions (as for the classic PAA mechanism).
Also in this case, the idea is to map all the SGE questions for individuating patterns and/or discovering topical areas that did not surfaced with the previous analyses.
For instance, in our case, an extremely important topical area comes out first thanks to the Follow Up questions analysis: mini painting.
AI questions analysis with Perplexity.ai
While waiting for how SGE will improve with Gemini, a better analysis of the potential conversations people can have with AI, hence for which to have our content being used and cited, can be done with Perplexity.ai.
The best way to use Perplexity.ai for this kind of study is using its copilot.
When using copilot, we can choose to interrogate the web as the source or to specifically use YouTube or Reddit as sources.
However, do not limit yourself to scraping the sequence of the questions, and note down the topic categorization that Perplexity.ai is asking you to segment the conversation:
And, particularly, the queries the copilot uses for searching:
The fourth secret hidden in plain sight: Images Search tags.
If you know me, you know how much I am stubborn about the importance of analyzing the Images Search tags for Entity Search analysis.
I consider them essential also for setting up a strategy able to target the most relevant search journeys because, if you have not noticed already, all the data I am retrieving with analyzing these less known/considered search features is meant for creating a database to use for doing Entity Search analysis.
Images Search tags present to us entities related to the entity object of our query (i.e.: “Stormtroopers” are an entity related to the entity “Star Wars: Legion) and attributions related to that main entity we are searching for information about (i.e.: 32 mm is an attribute of the Star Wars: Legion minis).
On average, we can go as deep as 4 levels when clicking on an Image Search tag.
Multiply each tag of the first level with all the tags of the following explorable levels, and you can understand what a wonderful and precise database of related entities and attributes we can retrieve from this analysis.
Topically.io is probably the only tool that can help you in this analysis, which otherwise would be painful:
However, as a practical tip for the use of this data in combination with all the others retrieved in the previous analyses, it is better to add the main entity along with the discovered ones in our database.
For instance, if “Stormtroopers” is an entity associated with “Star Wars: Legion”, then in our sheet we will note it down as “Star Wars: Legion Stormtroopers”.
Why? To better support the tools we will use later in disambiguating the specific nature of the entities.
The fifth secret hidden in plain sight: People Also Ask.
Well, the PAA feature is not a secret.
However, the way I recommend you use them now is surely less common: entity search analysis.
It is more normal to retrieve People Also Ask questions for creating content answering the questions themselves and, in the more pedestrian use of them, creating a single post for every single question (rolling eyes at 200 km speed).
Instead, use Alsoasked.com to retrieve up to 3 or 4 levels of connected People Also Ask questions, and add them to the database of queries you are summing up with analyzing the SERPs.
The sixth secret hidden in plain sight: the web pages ranking in the top 10.
This is the last analysis I recommend you do from an entity search perspective.
In fact, besides all the potential “ranking factors” at play, if a web document is presented in the top 10 by Google, that is because they are relevant for the entity implied in the query used for a search.
To facilitate our analysis, we can use the Google Sheet extension of WordLift, which analyzes the top 10 ranking search results for the query we ask to review, and for each search result page, it individuates what the entities present in the pages and their confidence level.
This is a relatively easy way to start setting up an entity prioritization in our work.
Ok… and what must we do with all the data we have retrieved?
It is now the time to do NER (Named Entity Recognition).
ChatGPT is probably the most useful tool we can use right now for Named Entity Recognition.
For this exercise, I retrieved a small set of data, so I could do NER directly in the conversation box of ChatGPT.
However, the ideal solution is to create your GPT with the right prompt’s commands and use the database of queries/topics/questions we have assembled with our analyses of the SERPs.
The first half of the prompt is this:
Extract the named entities from the following query phrases that belong to the “Star Wars”, “Star Wars: Legion” and “miniature painting” domains: [list of queries].
The second half is this:
Organize the named entities you have extracted in a table composed of two columns: the first one indicates the most relevant domain each named entity is part of and the second column indicates the named entity. Indicate, then, if a named entity is shared by more than 1 domain.
The result will be something like this:
Based on the (limited) dataset we have, we can easily see the main areas our website should create/optimize content for better targeting the entities it must be relevant for:
- Physical Component of the game.
- Factions/Units of the game.
- Roles in the game.
- Gameplay/Rules.
- Miniature Painting.
- Star Wars universe and lore.
If you are not into creating your own GPT, you can use the clustering function of a tool like Keyword Insights, which allows you to obtain something similar:
Moreover, that same function presents you with the data in a bubbles vision, which is not that different from the graphical vision of embedding clusters and how they relate to each other:
Ok… now that we have retrieved the data from the SERPs and then, used it for clustering entities and topics, what will we do?
We will finally work on the fundaments of our optimization strategy: taxonomy and ontology.
Taxonomy
A taxonomy is a method of organizing data. Taxonomy represents the formal structure of classes or types of objects within a domain. A taxonomy can also be a set of chosen terms used to retrieve content online (for example, a website’s taxonomy).
In other words, a taxonomy is a way to structure the content of a website and Google and AI – no matter how good it is in NLP – appreciate dealing with structured information.
Obviously, in our hypothetical ecommerce specialized on Star Wars: Legion, we already have a taxonomy, the one provided by Atomic Mass and Fantasy Flight Games, but it is standard, used by many other distributors like us, and it is not enough for targeting all the potential search journeys we have discovered thanks to our analyses, entity search and clustering.
Thanks to our previous work we can offer many other different navigation opportunities to our users and target all the possible search journeys and related query sets very naturally… and without needing to analyze dozens of lists of hundreds of thousands of unconnected keywords spat by classic SEO suite.
Our previous work suggested a taxonomy per faction type:
And another per Star Wars lore (in our case, Star Wars battles):
And a taxonomy per the basic nature of the game units:
Or per the typology of the game:
Or, to add another one, per material with which the minis are done:
- Hard Plastic.
- Resin (3D Printed).
- STL (Digital)
In other words, the faceted navigation of the product database that we should target and people search for real.
However, our work also offers us the opportunity to discover the informational content that we need to improve or create to target other potential search journeys inside Google and be meaningful for our potential customers.
And not simply to discover it, but also how to structure it in content hubs as it is, for instance, in the case of miniature painting:
The need to update the taxonomies.
We must remember, though, that taxonomies are not sculpted in marble. They evolve and change with time.
The best way to understand when and how to update our taxonomies is by using Google Trends.
We can use a tool like Keytrends.ai for facilitating and mostly automatizing the monitorization of Trends and for timely updating your website architecture.
Ontology
I am used to comparing taxonomy to a bidimensional vision of the entities of a domain, and ontology to a tridimensional view of those same entities.
An ontology, instead, provides schema information and axioms and/or rules for consistency.
Rather, taxonomies provide information about hierarchies of concepts and things.
We are talking about structured data and Schema.org, but – considering that it would be needed an entirely separate guide for discussing its best use – I will limit myself to this recommendation: do not use Structured Data as a synonym for Rich Results.
Rich Results are a nice useful output for some structured data, but the real value of structured data is on the parsing level, facilitating the machines to better understand:
- The meaning of what they are parsing.
- The semantic relations between different documents and resources present on a website and more (web semantics).
- The relation of all the entities presented on a website and how they relate with all the other entities existing on the web.
For this post, and of the virtual Star Wars: Legion ecommerce used for my exercise, it is now relevant to talk about the correct use of ontologies for structuring the unstructured, aka how to write texts that are consistent, relevant, easier to understand by machines (and humans) and easier to connect with other contents closer in the domain space is optimized for both the Knowledge and Shopping Graph.
I am sure many of you already understood that I am talking of Entity Salience and Context, something – once again – that is not new at all but that tends to be “forgotten”, even though it is explicitly presented by Google in its NLP documentation:
An entity represents a phrase in a text that is a known entity, such as a person, organization, or place.
An entity’s relevance score provides information about the importance or centrality of an entity throughout the text of the document.
Scores close to 0 are less important, while scores close to 1.0 are very important.
Let’s see an example.
Surviving the fall of the Galactic Empire, Moff Gideon soon takes command of a massive Imperial force in disarray, reorganizing it to counter the newly established New Republic.
Armed with the Mandalorian black lightsaber, Moff Gideon will find his fiercest opponents in Mando, Grogu, and the Mandalorians led by Bo-Katan Kryze.
This is a good basic text for every website related to Star Wars but – even if we can understand that it is about Moff Gideon – it is not able to let us understand the context into which we are presenting this bad guy of the post Galactic Empire period.
A better use of ontologies is needed to present better the context we are writing within.
If we continue the text with:
Moff Gideon appears in all three seasons of The Mandalorian that have been aired so far.
Played by cult actor Giancarlo Esposito (Breaking Bad), Moff Gideon has been the subject of many theories by fans of the series, such as his role in cloning Emperor Palpatine or his death at the end of the third season.
Now, it is clearer that we have prepared a short but meaningful character description for a website about TV series.
Instead, if we add this to the initial text:
Moff Gideon, like all Star Wars: Legion minis, comes in hard plastic. Easy to assemble, it faithfully reproduces the face of the actor Giancarlo Esposito who played him in The Mandalorian.
The mini is accompanied by 1 unit card, 1 upgrade card, 3 command cards, 1 token, and 1 instruction sheet. The Moff Gideon miniature comes not painted.
It is now obvious that we are talking of the Commander Unit “Moff Gideon” for the tabletop game “Star Wars: Legion”.
Ontologies, related products, and Visual Search.
Remember when I stressed the importance of analyzing and retrieving Images Search tags?
They are of great help for understanding the relations between the entities and their attributes, which means using them as datasets and applying an even simple embedding analysis, we can obtain valuable insights about two essential aspects for every type of ecommerce website:
- Related products.
- Visual shopping facets.
Let’s “forget” for a moment our virtual ecommerce selling Star Wars: Legion miniatures, and imagine we are a fashion brand specialized in selling armour, uniforms and apparel for the Galactic Empire’s army and fleet.
Then, we are very aware that visual shopping (and showrooming) is very common in fashion… therefore in our SEO strategy, we should prioritize the optimization of our catalogue of products for it, which translates to Google Lens.
With Lens, we know people can search for objects within an image.
Thanks to embeddings, visual components are transformed into numbers, and thanks to our previous entity search
analyses, we can easily see the closeness between products/objects even if they are taxonomically different.
Lens works for objects, so if it recognizes them, it highlights them so that searchers can select the object, which they find more interesting… even if it is not the one they were searching in the first moment.
Resuming, the entities “database” we have retrieved before, which we used for creating a better map of taxonomies and for optimizing the content of the website, now it can be also used for optimizing images for “related products” in Lens following these steps:
- Embedding analysis.
- Entity clustering.
- Neighborhood analysis.
Then, having used Image tags for entity search, we can also know what attributes – for instance, the colours for the stormtroopers’ armour – are more recurring hence, what visual facets we can present as alternatives in Merchant feed and Product structured data and, voilà: people search for a stormtrooper armour using Lens and then, thanks to Multisearch in Lens, precise they want to know if a pink coloured armour is available, and we will have a product image to show them.
For better integrating the ontology research insights in a practical and actionable way, I must recommend Wordlift once more and, more specifically, its Product Knowledge Graph builder.
The seventh secret hidden in plain sight: the product Knowledge panel.
Back in January 2020, Google rolled out the “Popular Products” search feature, the first step of introducing organic shopping (or merchant) in the Search. Since then the presence of the Shopping Graph in the ecosystem of Google has become overwhelming.
One of the features it populates the SERPs with is the Product Knowledge Panel, which once offered a navigation per tab but that, now, is almost blended in the SERP with elements like:
- The main product view.
- Top Considerations.
- Common uses.
- Theme and Design.
- User Reviews.
- Q&A.
- Discussions and Forums.
- Expert Reviews.
Just by viewing the list of the elements composing a Product Knowledge panel, we can start understanding the importance of Merchant feed, of a video strategy, and of the two E of E-E-A-T: Experience and Expertise.
However, if we dedicate more time to analyzing them, we will discover important suggestions for better assessing our SEO strategy.
The main product view and the Details block
People usually think that these two sections of the Product Knowledge Panel are based only on the Merchant feed, and that is an incorrect belief.
These sections use data from the Merchant feed AND the Manufacturer feed.
The two feeds are slightly different, as explained in the Manufacturer Center vs. Merchant Center feeds help page:
A Merchant Center feed is primarily used to help a merchant specify details that aid in the sale of the product. […] Merchant Center accounts are available to any retailer. […]
A Manufacturer Center feed is primarily used by branded manufacturers to share detailed and rich product information such as product titles, descriptions, images, key features, YouTube videos, and others that are not captured in a Merchant Center feed. The data submitted into Manufacturer Center is used to enrich Google’s overall product catalog. Manufacturer Center is only available to manufacturers, brand owners, and brand licensors, regardless of whether they sell directly to consumers. […]
TL;DR: if you are both a manufacturer and a merchant, use both feeds so to have richer product knowledge panel pages for your products, hence, making them more able to convert.
Top Considerations, Critic Reviews, Videos and Common Uses.
All these sections of the Product Knowledge panel are related to Expertise and indicate to us that we must include outreach and promotion from the very beginning, and not just to get links (which are still important, but to create brand and product awareness.
It is important to analyze who writes about our competitors and whose content is used by Google as a source of EXPERT information.
Once we have retrieved these insights, we must analyze the true influence of these expert sources using tools like any backlinks analysis tool, Sparktoro (particularly good for discovering websites and real influencers associations) Vidiq.com and, finally, set up an outreach calendar.
However, apart from looking for classic reviews on experts’ websites, a smarter use of influencers would be to create co-marketing actions like asking them to create short product reviews that we will publish on our PDPs and contributing in the informational section as, for instance, with videos about how to paint one miniature of choice of Star Wars: Legion.
Then, it is important to know that:
- Top Considerations and Expert Reviews are not affected by QDF.
- In some cases, Google uses the comments present in the cited source.
Finally, in the case of videos, we should also monitor the presence of YouTube Shorts, also because they are quite commonly found in Perspectives.
User Reviews and Q&A.
The reviews can be from “Google”, i.e. extracted from reviews made directly on Google Merchant (paid and/or organic), but they can also be reviews indexed by websites.
This means that we must set up a strong strategy for nurturing customers by writing a review for the product(s) they bought, presenting Google Shopping as the main means for doing it but offering as an alternative option our website, and so implement “Experience-based” content on our website and product pages.
Secondly, the User Reviews feature suggests that we monitor the reviews about our products that customers leave on affiliates’ or distributors’ websites, and eventually engage with both positive and negative reviews.
In other words, we should use for ecommerce users’ reviews the same strategy and tactics local SEOs are using for local businesses.
The Q&A feature, then, is useful for individuating the most common questions people ask about our product and, therefore, the ones we should answer directly in our PDPs.
Discussions and Forums.
Reddit Reddit everywhere.
We saw it already when exploring the potential search journeys inside Google, and we see it now also as the main content feeding the Discussions and Forums section of the Product Knowledge panel. This feature is not yet present in every Product Knowledge Graph, but you can see an example here below for a different kind of product than a tabletop game:
As said before for the Users Reviews, the Discussions and Forums is a clear suggestion for us to monitor the forum space. However, we should not limit ourselves to the classic Reddit and Quora, but also to less known but important for our niche forums, and Google surfaces in this feature.
But we should not limit ourselves to monitoring them.
First, we should individuate who are the persons that can be defined as more influential in those forums and outreach them – as we did for professional journalists for the Reviews – to create content for our website.
Secondly, we should engage with the members of the forum, answering questions, solving doubts, asking for feedback, launching exclusive previews of new products, et al.
The ultimate secret…
Compliments! You have reached the end of this long post!
And doing so, you were able to discover how Google itself can be used as a tool to assess and optimize our SEO strategy and how to create a strategy based on solid foundations of taxonomy and ontology, as well as observation and collaboration with all those actors external to our organization (customers, specialized journalists and creators) allows us to have a site that is largely immune to things like Core Updates, Helpful Content and Review Updates, and capable of being visible to our potential user/customer along the greatest number of potential Search Journeys within the Google ecosystem.
But the secret implicit in this post and the exercise, I created is the profound knowledge I have of the niche of tabletop games, miniature painting and the communities of players and enthusiasts of these leisure options.
Without that knowledge, a knowledge that we should try to obtain for all the market niches in which our clients operate, we cannot obtain the creative plus based on unexpected associations and lateral thinking, which is what makes our work as SEOs truly useful.
May the Force be with you, always.
