Topical Authority and Semantic SEO will be discussed more frequently with the concept of Structured Search Engine in the coming years. In this article, I will explain how I used these techniques to grow monthly organic traffic from 10,000 to 200,000+ on Interingilizce.com in just 5 months.
Last 5 Months. Graphic of GetWordly.com. Project had started 6 months ago.
When you read this SEO case study, some things may seem strange to you because they’re a bit extraordinary. Therefore, we will need to highlight the not so familiar concept of semantic SEO.
Comparing the Last Three Months and the First Three Months from GetWordly.com’s GSC Organic Traffic Performance Report.
Background of the Semantic SEO Case Study and Projects
During this SEO case study, none of the following methods were used, and all the important visible SEO elements below were intentionally excluded from the projects.
Brand Power and Branding
Technical SEO (That’s right: I didn’t use it.)
Quality Web Page Layout and Design
A Healthy Server
In summary, anything we normally say is good great practice in a normal SEO case study was not carried out on this website. Therefore, you may think that the success I have had is a fluke. But you would be wrong: with the same methodology, I have created 4 separate SEO case studies and success stories in the last 5 months.
As an example, “I even had lots of “poor URLs on desktop”.
This article is an executive summary of my methodology and also a manifesto for semantic SEO and the true values of SEO theories with free analytical thinking.
This means I have decided not to cover related topics and experiences, such as:
Minor Google updates
The reason I lost all featured snippets in one day
How our server collapsed from excessive traffic
The effect of server failure on search engine algorithms.
With that in mind, let me present the methodology.
4 Different SEO Projects, Same Methodology and Same Success with Semantic SEO
In this article, I will focus on Interingilizce.com with a simple semantic SEO executive summary. But, to reduce the questions you may have and guarantee that “SEO Theories and Google Patents” has a concrete and actionable value (thanks to Bill Slawski), I want to summarize the 4 different SEO projects’ journeys with only their results.
If you want to read all the heavy SEO terms, theories, patents, and practical details about these SEO case studies and projects, I recommend you read my Topical Authority, Coverage, and Contextual Hierarchy for SEO article. (It’s a long one and has more than than 14,000 words.)
“We were aiming for success in a short time with a quick effect in our new projects. And we surpassed our targets in our new project.”
Head of Marketing, KonuşarakÖğren
First Project: 1100% Organic Traffic Increase in 5 Months, Interingilizce.com
Interingilizce.com is actually 2 years old, but there was neither content on the website nor organic traffic. I have increased the Org. Traffic by 1100% in 145 days. From 10,000 to 200,000 clicks. Since you already saw the “last 6 months organic traffic” result in the intro section, you can see the “May 2020” and “November 2020” comparison below.
I know, actually, it is way much more than just the 1100% organic traffic increase, but I can’t use the “%” sign for this kind of increase to make you imagine something concrete enough.
Below, you will see Interingilizce.com’s Last 3 Months Comparison graphic for organic traffic.
And, last 6 Months Organic Traffic Graphic for Interingilizce.com.
You might confuse it with GetWordly.com’s GSC Report, but they are different. Since, I use the same methodology, they had the same reflexes against Google algorithms and updates.
Second Project: From 0 to 330,000 Monthly Organic Traffic, and 10,000+ Clicks Per Day: GetWordly.com
GetWordly.com is the second site, the website was pure in terms of organic visibility and history. On GetWordly.com, I reached the level of 11,000 organic clicks per day and 330,000 organic clicks per month in just 6 months.
You will see GetWordly.com’s last 6 months organic traffic graphic from Google Search Console below.
Below, you will see GetWordly.com’s Ahrefs graphic.
And, below, you will see the SEMRush Graphic for GetWordly.com.
Third Project: 600% Organic Traffic Increase in 5 Months: Focused on Azerbaijan
Third website (Name not disclosed yet) saw 600% growth achieved. Within 5 months, monthly organic traffic increased to 70,000 from 10,000. (The traffic volume is lower than the previous examples as the third website only targets Azerbaijan.)
Below, you will see the last 6 months’ graphic for the third site’s GSC Graphic.
And, below you will see the last 7 months of the “dictionary section”.
You can see that “in one day” the “dictionary section” of the website has lost all of its traffic. We had changed nothing. All the difference was coming from Google’s own internal systems. Since Google started to fully rely on “Deep Learning” and “Machine Learning”, I know that they also evaluate “page layout” according to the “general feedback” that they collect from the web.
Last 3 Months Comparison for the Third Project.
Thus, we just changed the layout of pages, with some “background color changes” and just “order of the page elements”, and everything has gone back to normal, in two days.
While this part of this SEO case study and experiment is not about semantic SEO, as a Holistic SEO, I just wanted to say that sometimes you should focus on “different things”. Below, you will see “Website Representation Vectors” from Google Patents.
Website Representation Vectors can be used to understand “possible user satisfaction” after a click by a search engine.
Fourth Project: 400% Organic Traffic Growth – Focused on Arabic Countries
The fourth website (Name Not Yet Disclosed) targets the entire Arab World. Since my main focus was not on the website (since I don’t speak Arabic), only 400% growth was achieved.
Below, you will see the last 6 months’ graphic for the fourth site’s GSC Graphic.
In my opinion, for this website, we didn’t have enough traffic increase according to the power of semantic SEO. That’s because of a lack of Arabic language skills. Technical SEO is “language-agnostic” while semantic SEO is heavily bound to the nature of words, terms, concepts, and languages.
As a final proof of semantic SEO, you will see the 3 Months GSC Organic Traffic Performance Comparison for the fourth project.
A Brief Explanation of the Semantic Web, Semantic Search, and Topical Authority
The semantic web is the state of the information organization on the web. The semantic web uses two basic elements that come from the nature of the human brain’s and the universe: taxonomy and ontology.
Taxonomy is coming from “taxis” + “nomia”, which means “arrangement of things”. Ontology comes from “ont” + “logy” means “essence of things”. Both are means of defining entities by classifying them into groups and categories. Together, taxonomy and ontology make up the semantic web.
Over the past ten years, Google has created several initiatives moving towards a semantic web.
In 2011, Google announced the “Structured Search Engine” for structuring the information on the web.
And, in May 2012, they launched Knowledge Graph for better understanding information about real-world entities.
In 2019, they launched BERT, a model for a better understanding of the relations between words, concepts, and entities in human language and perception.
All of these processes created the semantic web, semantic search, Google as a semantic search engine, and consequently semantic SEO.
Word count evaluation of GetWordly.com in OnCrawl. Most of the content from this project includes more details and information than its competitors.
What do Topical Authority and Topical Coverage stand for in this Context?
In a semantic and organized web, every source of information has a different degree of coverage for different topics. Things or entities are connected to each other through their shared attributes. These attributes represent “Ontology”. Things also connect to each other within a classification hierarchy. This hierarchy represents the “Taxonomy”. To be an authority for a topic in the eyes of a semantic search engine, a source should cover a thing’s different attributes within different contexts. It must also reference similar things, and things in parent and child categories.
Creating a content network for every “sub-topic”, for every possible question, within contextual relevancy and hierarchy with logical internal links and anchor texts is the key for these SEO case studies.
As OnCrawl says, I didn’t use a healthy sitemap during this SEO case study. Site here is again GetWordly.com
Topical Authority and Topical Coverage can be acquired with the most detailed, entity-oriented, and semantically organized, comprehensive content network. Every successful piece of content increases the chance of success of other content for the connected entities and related queries.
To keep things brief, I will move on to the “do” and “don’t” section. If you need more details or understanding on these topics, I recommend you read the article I mentioned before.
What Should You Do to Implement Semantic SEO?
To understand the concept of semantic SEO fully, you need to understand why a search engine needs the web to be semantic. Especially with the dominance of machine learning-based search engine ranking systems instead of rule-based search engine ranking systems and the use of natural language processing & understanding technologies, this need has increased even more. So approach these concepts through the eyes of a search engine to understand the suggestions below.
Since I didn’t use any kind of “header” or “footer” menu, most of the pages were really really in deep.
1- Create a Topical Map Before Starting Publishing the First Article
Remember ontology and taxonomy? You should check Google’s Knowledge Graph because for Google, things can connect to each other in a different way than according to a Dictionary or Encyclopedia. Google uses the web and information provided by engineers for its entity recognition and contextual vector calculations.
Thus, you also should check SERP to see which entity has been connected to which and in what way for which queries…
Inrank is OnCrawl’s “Internal PageRank” Distribution formula that is being inspired by Google’s first and original PageRank formula. And we see that Inrank has fallen for pages that are far from the homepage.
This might be a little bit exhausting but after a while, you will see how Google thinks, acts, and connects things to each other. There are some quick-practices you can do to check a niche and query group for creating a topical map.
Crawl your competitors’ sitemaps to understand their topical maps.
Pull Google Trends related queries and related topics.
Collect data from autocomplete and search suggestions.
Notice how your competitors are connecting content hubs.
Use Google Knowledge Graph to pull relevant entities.
Use non-web resources to view the properties of entities and their hierarchies and connections.
The last item is also important to become a resource that provides original and authoritative information for a search engine’s knowledge base.
P.S: Also use other search engines. I especially recommend you check Swisscows to understand a semantic search engine’s nature. Do not solely focus on Google to understand semantic search.
This is a Topical Map (Topical Graph) example from Inlinks.net. Since, this is a unique technology and InLinks is the only SEO Company focused on semantic SEO, I recommend you check their technology.
2- Determine the Link Count per Page
In all of these SEO case studies and achievements, the total number of links on each web page was at most 15.
Most of these links were in the main content, with relevant and natural anchor texts. I didn’t use a footer or header menu. This contradicts traditional recommendations in technical SEO. I had to accept that and I am not saying that you should use no more than 15 links per web page. I am saying that you should keep the relevant and contextual links within the main content, and try to make search engines focus on those links.
You can use the following items to determine the correct number of internal links that should appear on a web page:
Industry standards for internal link count to understand the minimum and maximum values.
Number of Named Entities in the content
The number of contexts for Named Entities
The “granularity” level of the content
A maximum of 1 link in each heading section
Linking entities belonging to the same type to their respective pages if they are in “list format”.
But, I will briefly tell you that I never use an anchor text more than three times for a web page within the main content. In other words, for the fourth time, I recommend you add more words or change some words within the anchor texts.
I also have some other types of rules for anchor texts:
I never use text in a page’s first paragraph as anchor text in links to that page.
I never use the first word of any paragraph on the page as anchor text to that page.
If I link an article to another article from a different context or a side-topic, I always use one of the last heading’s paragraphs (Google calls this type of connection “Supplementary Content”.)
I always check competitor’s anchor texts for a specific article internally and externally.
I always try to use synonyms for a topic when creating anchor texts.
I always check whether the “anchor text” exists in the targeted web page’s content and link source’s related heading text or not.
P.S: I am not saying that you have to do this to do successful semantic SEO. These are just a few examples of guidelines I followed when I obtained these outcomes. If you find some examples that are not compliant with these guidelines, it is probably because of my lovely authors.
OnCrawl’s Inrank Flow Distribution for different URL Categories. Since I didn’t categorize anything on these projects, it is mostly “green” and means that the most important part is the blog.
4- Determine the Contextual Vectors
This terminology can be a little bit “scratchy” for your ears, again. This is a term from Google Patents, for me. Contextual vectors, contextual domains, contextual phrases… There are lots of things to dig in Google Patents (another thank you to Bill Slawski, our educator).
Simply put, contextual vectors are the signals to determine a content’s angle. A topic can be “earthquake” and a context can be “comparing earthquakes”, “guessing earthquakes”, or “chronology of earthquake”.
For instance, “apple” (fruit) is an entity and also a topic, and Healthline has more than 265 articles for just “apple”. Benefits of apples, nutrition of apples, types of apples, apple trees (basically, a different entity and topic, but it is close enough.)
So, in this context, all of these sites were from the second language education industry. “English Learning” is the main topic; learning English from games, videos, movies, songs, friends… are different contexts.
To create more contextual connections, I always try to fill the gaps between different topics and entities within them with the help of different types of pillar cluster contents. I also recommend you to read about Google’s contextual vectors and knowledge domains from their patents.
OnCrawl’s internal link status table. There was not a clear “internal link” status code and crawling path for the search engines.
5- Determine the Content Count be Written and Published
Content count is not a ranking factor. Actually, telling more things with less content with more comprehensive and authoritative articles is better for many aspects such as crawl budget, PageRank distribution, backlink dilution, or cannibalization problems.
But, content count is important to plan the process. Because, you will need to know how many authors you will need, or how many articles you will publish per day or per week. I didn’t include lots of SEO terms in this executive summary such as content publication and content update frequency… But, even if you determine the topics, contents, contexts, entities, you still don’t know how much content you will need. Sometimes Google prefers sites that show different contexts for a topic on the same page, but in other cases Google prefers to see different contexts on different pages.
Average Heading Count per heading level on the web pages are above. Heading count per heading type can reveal the content detail, length and granularity level for an SEO.
To know the exact count for the content/article, examining the Google SERP types, competitors’ content network’s shape is important. This is also important for the budget of the project. If you tell your customer that you just need 120 pieces of content but later, you realize that you actually need 180 pieces of content, it is a serious problem for trust.
And, for any SEO success story, clear communication is a must.
From the number of H2, you may understand the detail level on a web page if the source doesn’t use “gibberish sentences” for their content structure.
6- Determine the URL Categories and Hierarchy
URL categories were not used in any of the SEO case studies here. However, this does not mean that URL categories and corresponding breadcrumbs are not beneficial for semantic SEO. Keeping similar content in the same folder in your URL path makes it easier for a search engine to understand a website. It also provides tips for users and facilitates in-site navigation.
If that’s the case, why didn’t I use it? For the same two reasons I am not using technical SEO: because of time constraints and because I want to run a future SEO experiment.
Problematic and Managed Duplications exist site-wide for these SEO Projects.
7- Create a Topical Hierarchy with Contextual Vectors by Adjusting it with URL Hierarchy
Sub-topics have been confirmed to be used by Google during January 2020 but actually, Google has mentioned them as “Neural Nets” or “Neural Networks” before. In the Google Developers’ YouTube channel, they also showed a nice summary about how topics are connected to each other within a hierarchy and logic. That’s why again, taxonomy and ontology are key for semantic SEO.
But, what does “creating a Topical Hierarchy with Contextual Vectors” mean? It means that every topic should have been processed with every possible context and related entities by being grouped with a logical URL structure.
This will lead the search engine to give a source better topical authority and expertise thanks to a more granular and detailed information architecture.
We also have “canonical conflicts”.
8- Adjust Heading Vectors
Heading Vectors… Another term that is foreign to many ears. Heading vectors are actually the order of the headings as a signal for determining the main angle and topic of the content. According to the Google Quality Rater Guidelines, content is viewed as having three different sections: “Main Content”, “Ads”, and “Supplementary Content”.
We all know that Google gives more weight to the content above the fold or in the “upper section” of the article. That’s why the queries on the upper section of the content always have a better rank than queries on the bottom section. To Google, the bottom section actually represents the “supplementary content.”
A representation of Google’s methodology for Contextual Answer Passage Score Calculating via Heading Vectors.
That’s why using contextual relevancy and logic within the heading hierarchy is important. Simply put, here are some basic guidelines for heading vectors in my perspective on semantic SEO:
9- Connect Related Entities for a Topic within a Context
Connecting entities and entity associations are close terms to each other. Entity association can be done by search engines based on the entities’ attributes and also through how queries are phrased for a possible search intent.
Connecting entities and associating entities to each other within a context is a practical application of ontology. For instance, in the context of the industry of these SEO projects which is “English Learning”, for the topic of “Phrasal Verbs”, you can also use “Irregular Verbs”, “Most-used Verbs”, “Useful verbs for lawyers”, “Etymology of verbs of Latin origin”, “Less Known Verbs” which can be connected to each other.
SEO related HTML Tags should be optimized in terms of length and terms within them. In these projects, I didn’t implement these, as you might see.
All those contexts actually focus on “verbs in English”. They are all related to “Grammar Rules”, “Sentence Examples”, “Pronunciation” and “Different Tenses”. You can detail, structure, categorize and connect all these contexts and entities to each other.
After you basically cover every possible context for a topic and all related entities, a semantic search engine doesn’t have any other chance besides choosing you as a reliable source for possible search intents for these.
10- Generate Questions and Answers for Possible Search Intents
“Generating Questions from Answers”… Another Google Patent. But, since this article is long enough already, I won’t go into detail. Basically, a search engine generates questions from contents on the web and matches these questions with queries with query rewriting. And it uses these questions to close the possible content gaps for possible search intents on the web.
That’s why I tell you to process every entity with every context while connecting them to each other. However, you should also know what information extraction is. Information extraction is pulling out the important facts and definitive connections about concepts from a document. Thanks to information extraction, a Search Engine can understand which questions can be answered from a document or which facts can be understood. Information extraction can even be used to create a knowledge graph between entities and their attributes, and used for generating related questions.
Generating Related Questions for Search Queries
You shouldn’t just focus on SEARCH VOLUME! It might even be that no one has asked this question before. And even the search engine might not know the answer to this question. But if this unique information is useful for defining the attributes of entities within the topic, generate and answer these questions, and be a unique source of information for the web and for search engines in your niche.
11- Find Information Gap Instead of Keyword Gap
First, read the quote below.
“An information gain score for a given document is indicative of additional information that is included in the document beyond information contained in documents that were previously viewed by the user.”
Source: Patent “Contextual Estimation Of Link Information Gain”
We all know that even as recently as 2020, “15% of queries everyday are new and Google uses RankBrain to match these queries with possible search intents and new documents”. Google also always searches for unique information and answers for possible future queries from its users. Provide unique information, and try to contain less known “terms, related information, questions, studies, persons, places, events and suggestions”.
So, “longer content” or “keywords” are not the key for these SEO case studies. “More information” and “unique questions” and “unique connections” are the keys. Every piece of content for these projects has unique headings that are not even related to the search volume and even users may not be aware of them.
Below, you will see another Google Patent to show the contextual relevance for augmented queries and possible related search activities.
Augmented Search Queries Using Knowledge Graph Information
“Including every related entity with their contextual connections while explaining their core” can be seen also here with its importance.
12- Stop Caring About Keyword Volume or Difficulty
I’ve already shared my opinion about keyword volume. During these four projects, I saw myself mostly as creating a quality and understandable “educational book”.
At the beginning of the project, authoritative competitors with tons of backlinks did not intimidate me.
I didn’t care about third party metrics like keyword difficulty.
The historical data and brand strength of the competitors did not scare me.
And finally, I avoided resorting to “I just used Google Search Console to show my client the latest situation of projects”. Except to review Google’s reactions, I didn’t get into GSC.
When writing articles, if a subtopic is required in a topic’s semantic structure, it should be written. It should be written even if the search volume is “0”. It should be written even if the keyword difficulty is 100.
There is another important point here.
If you want to rank first in the SERP for a “phrase”, you need to include all relevant phrases, and every detail in all relevant topical graphs. In other words, it is not possible with semantic SEO to see an increase in rankings in queries related to that topic without fully processing each related topic.
Word Count Evaluation by page depth. The older the content gets, the page click depth increases on this example since we don’t use a standard internal navigation. But even in the 10th depth, we have stronger content than our competitors. This encourages Google to look further and deeper.
13- Focus on Topical Coverage and Authority with Historical Data
Topical Authority = Topical Coverage * Historical Data
That’s why in every graphic that I show you, you see a “rapid growth” after a certain amount of time. And since I use natural language processing and understanding, most of this initial rapid organic traffic growth with a wave shape comes from featured snippets.
If you can take featured snippets for a topic, it means that you have started to become an authoritative source with an easy-to-understand content structure for the search engine.
N-Gram Analysis is one of the best features of OnCrawl for me. And it is a unique feature for an SEO Crawler. You can see “6-gram Analysis as site-wide” above. “Learning and teaching English” (İngilizce Öğrenmek ve Öğretmek) is the main theme of this entity as it can be seen.
14- Use Natural Language Processing Terms with Your Authors to Determine the Best Possible Sentence Structure and Content Format
Educate your authors.
Show them how Google uses NLP and NLU. Teach them what “Part of Speech Tags” are or what “Named Entity Recognition” and “Named Entity Linking” are. Use N-Grams, Skip-Grams, Word2Vec with concrete results and practices to make them understand machine-run text analysis. Show them how the Google Knowledge Graph works.
Teach them what Neural Matching or Entity Type Matching is. Show them their mistakes over Google Docs while revising their content, then show them the organic traffic growth with featured snippets.
I sometimes call this “Featured Snippet Oriented Content Marketing”. We all know that Google simply uses NLP models to understand the content and to be honest, beating even the most authoritative competitors is easier than it was in 2011, thanks to Google’s NLP-based algorithms.
And, as I said at the beginning of this Semantic SEO Manifesto, I lost all featured snippets in one day. This occured when Google performed a minor update that decreased the percentage of featured snippets by 4% in one day. At the same time, our server has been down due to excessive traffic. Additionally, to create a “reranking demand”, I had updated lots of content while continuing to publish new pieces of content.
How Google tries to understand whether a content is opinion-based or fact-based. Patent name: Machine learning to identify opinions in documents.
Below, you can find some of my rules for my authors:
Never give your opinion in an article.
Never use an “everyday language” in the articles.
Don’t use analogies.
Don’t use unnecessary words.
Content should be short as possible and long as necessary.
Always use short sentences instead of long sentences.
Always give answers directly and precisely.
Always use a “source” as an authority before giving a statement.
And, as a not big surprise, I didn’t use “alt tags” for images. But, since this is also an accessibility issue, even if it is for an SEO experiment, leaving a site in an inaccessible situation doesn’t feel right. Imagine that for an experiment, you remove the wheelchair entrance to a store. The same principle applies to websites.
Sometimes it is difficult to get your authors to follow these rules, and I am not saying that you must follow them.
I am saying that I (mostly) followed these rules during these SEO projects.
15- Educate Your Customer instead of Keeping them in the Dark
I see that most of the SEO agencies do not explain SEO’s subtle sides to their customers. The main reason for this lies the basic side of their business model. SEO is a business model based on the subscription economy model. It means that customers should continue to buy the service.
But you can also use the “IKEA Effect” for your business. The IKEA Effect is when you make customers contribute to the work, and it makes them love you more. As an owner of a “one-man-one-desk” company, I use my customers’ team for our SEO projects. In other words, I don’t need an SEO team for myself because I already have multiple teams that I educate.
OnCrawl’s response size report according to the click depth. Having a “bad design” can also be useful if you want to have “smaller response sizes”. It is not an intentional situation but it is a natural outcome.
And when the customer starts to understand, to learn from you and to work hands-on on their own project, they start to love SEO and they feel the IKEA Effect. They give more value to the SEO project through self-association and effort justification.
Thus, they will listen to you easier, and sharing the “real know-how” won’t jeopardize your business. Instead, it will make things easier. I even educate interns or sometimes enter job interviews for my customers’ job applicants, because their future team member is also my team member.
OnCrawl’s response time report. After a server crash due to the organic traffic increase, we have bought a new server. That’s why this crawl report shows better response timing. The other side of reality can be seen below.
During the SEO case study and experiment, even the most important thing was not enough: the server. A screenshot from Interingilizce.com’s Server Connectivity report from Google Search Console’s Crawl Stats Report.
As an entrepreneur, this is my own model that is born from my own conditions. I hope this also fits for you as in other articles. Semantic SEO relies on patience, costly authors, highly theoric SEO terms, and content engineering with algorithmic knowledge. Without educating the customer, it might not be possible to convince the customer to follow you on this road.
And when you do convince them, they are happier to work with you, even when you propose theories that sound new or different to them:
“Koray likes to apply new things to our projects in a very short time. We got huge spikes in our web sites with his strategy. We like his enthusiasm. If you don’t work with him, you haven’t seen an advanced SEO theory glossary.”
Owner of the KonuşarakÖğren
Last Thoughts on Semantic SEO
While writing this guide for this SEO case study with four different SEO projects, I have tried to keep things simple as much as possible. And I have told everything with complete honesty. If you can endure long theoric articles with a deep analytical analysis for SEO, I recommend you read the article that I recommended in the beginning that explains more than 40 different lesser-known SEO terms to understand everything behind this methodology.
As you have seen, in order to focus on an initial, rapid gain in traffic, I neglected a lot of SEO improvements for these sites. Working on these different technical SEO elements that you have seen in screenshots throughout the article will also improve traffic even more, but I wanted to be able to clearly show you the effects of semantic SEO alone.
Thanks to deep learning and machine learning, semantic SEO will soon become a more popular strategy. And I believe that technical SEO and branding will give more power to the SEOs who give value to the theoretical side of SEO and who try to protect their holistic approach.
See you in the future SEO case studies and experiments.
Koray Tuğberk GÜBÜR is a Holistic SEO Expert. He examines the Google Algorithm Changes by watching every inch of SERP. His main expertise is E-A-T Building and Technical SEO along with digital marketing management.