E-commerce SEO with Semantics: A Race between Entity Connections

hello everyone welcome to the series episode in this case will actually focus on one of the websites that I have explained three times already so this is one of the oldest projects that I am still active on but this activity is not that uh intense let's say it's mostly like a kind of strategical uh consulty work to determine a kinds of direction for the project I have announced this project three times with the k e that you just see here with certain type of updates and three times actually we doubl Tropic and it is actually

about entity oriented search people ask me sometimes what do you mean with the entity oriented search or entity oriented SEO basically it is meaning based SEO by connecting and associating entities with each other based on certain type of attributes when anyone actually searches any kinds of query the thir first thing that actually search engine is trying to do is understand your topic of the query than the main entity and attribute and possible values for this query today we in the results for this specific project because at the moment it is actually at the alltime High

version we will also explain why it has been affected from this section we will examine the Coss STS one more time and we'll examine also some of the problems in the website while also showing you some of the patterns to understand the product search and ratings for product search results as well so a SEO actually is designed for introduction of the advanced concepts for the seos and the business owners around the globe and this is a kind of community service from my point of view for my own community and with that say I say strong

strongly suggested to watch entire blue had video playlist because I don't repeat the same suggestions two times so these will be uh in this video this information that you see in the case it won't be actually answered one more time and you can see why these drops happen why these increases happen as well if you read the rest of the case study it will be really helpful for you so this is the search Atlas owned by the manic Bahan and I believe he is a good person and late I start to integrate some of the

websites of this system as you can see it directly makes a kinds of keyword ranking analysis in an aggregated way based on the search console API data and we see that for the 33,000 of the keywords we see an increase compared to the uh let's say last 28 days and we see also that uh declined ones also exist in this area too and later the traffic impression and everything is actually keep increasing since the 15 of August core update it is slightly being still balanced after the some unofficial updates as well the biggest problem in

the website is actually coming from that this section because if you have too many zero click uh Pages it means that you have a problem in your website so 33% of the website doesn't get that doesn't get that much clicks and also 57% of them also are close to zero as well so why this happens let's check it together bam the top and we right here let's say the piure print or something like that press means actually little little press is a very famous book if you look at here actually the website also became by

the way one of the biggest and largest e-commerce websites in turkey and here too you will see that same book has been given like 600 different versions you will see that there are different type of little princes from different type of publication centers from different also sellers different versions for adults for children and Etc normally all these variations actually should be going inside a single page while consolidating your actually ranking signals this is one of the biggest problems on in the source and this one if you go to the Page's report you will see that

it has like closer to 1.5 million uh not index pages and not everything in the site also has been indexed currently too as you can see the reason for that is most pages are actually very close to each other they don't have that much value response time is more than five times more than what I actually suggest and the good thing here is that this was way much worse right now it's better HTML section is also heavily better too but refresh or Discovery hits are not as it should be so it means that the source

actually still needs more publication frequency or update frequency the problem here is that for this type of large websites even if you update some certain sections it won't be visible to the search engine that's why it is better to give actually your uh your lify quality improve instuments and refresh freshness signals from the most important and popular books and book categories and aor profiles as much as possible in the topical map explanation here I explain how I have used aors together with genas to be able to increase the overall relevance by by uh empowering some

of the query templates so I want be diving that section again so to solve this issue if you have 600 different versions of the same book as six different URLs what the search CH will be doing is we will be sharing the ranking signals and all of the me will be ranking for the different different times of the day or times of the year and the week or the month so for these type of situations if you're able to increase actually your just the technical isue improvements and internal Le distribution at least you can directly

signal search engine that the this is the main version and the canonicalization can be performed at the behind of the scenes from the search engine point of view due to the business model you won't be able to usually theindex these Pages or you won't be able to remove these pages because your CMS doesn't allow you to do that and if you remove it then you will be having some heart problem with your customers or your partners business partners as well that's why SEO is a kinds of art between the business and the SEO to find

the best possible compromising or best possible balance to keep everyone happy Partners clients and uh also customers and also search engine if if all of them are happy then actually you are really helping the business as well so the result wise for the last three months I can tell that actually uh despite the impressions are slightly lower the clicks are still keep increasing and to be able to break this specific Source back actually for the most important books and most important authors the and also most important genes and subjects as well subjects that are not

related to directly the books or authors like timelines historical timelines eras or events are also evated and based on that also these specific subjects actually kept being processed so the first we go from the books then the authors then the genes then we go actually Beyond them like just for the timelines or historical events then we connect these books and authors to each other based on these events or based on these uh let's say subjects in that specific timeline as much as possible which means we expand the topical map further while doing that the C

STS have been actually improved really a good way because most of the unnecessary URLs are deleted and removed from the website through the last two years and thanks to that actually the website came into the alltime high uh let's say situation and right now it cats actually millions of collects including actually Google discover and other areas when it comes to other suggestion for blue head context it might be good to mention since you are a really big or large uh business here you're able to actually use official websites of the aors to be linked back

which means you publish a review for with them du the interview with the auor then the auor shares it and links you back you're able to use also lot of press releases and statistics uh to be about able to also linked back as well so we have used actually some of the let's say page rank acquisition meals here and since we have a large director of books and uh authors you're able to get also links from Wikipedia as well but as you can see 2. 700,000 referring domains is not that big either the important thing

here is that some of these links are actually coming directly uh from important new sources that section is good but if you look at uh let's say referring mean numbers it is increasing but most of the time this increase is correlated with your traffic increase I can tell because once you are more less say you have more exposure it will also affect the amount of the websites that see you and directly link you as well but but most of these domains are not that alterative you can also see it even from the numbers in this

area too so another advantage of this Source I always touch it we have lots of uh good amount of actually branded seource demand but none of the Branded search demands including page rank including refering domains including uh brand search demand is not even half of the competitors the computors are actually being searched way much more like three times or four times more than BM book.com or their page Rank and the referring domain numbers are also way much higher as well despite that we are able to actually outrank them by reading the algorithms or explaining the

internal substruction of the website despite the the CMS is not in the best conditions so when it comes to this section I would like to show you some of the products related actually uh let's say product search patterns that I have filtered earlier some of these are actually from important Engineers but they are always connected to each other for instance this is about product recommendations based on collaborative filtering of seller products this is about finding the same product across different type of Sellers and if multiple sellers have the same book like this one for instance

format hard cover publication date Etc and if you're able to get actually the connections how they recommend these products to each other by performing what type of actual checks you will also be able to create a better recommendation inal and you can also increase your relevance further as well here you will see multiple contexts uh for these type of products or their suggestions I would suggest to check this another one is uh from Linda probably from Vietnam a vietnames I'm not able to read this uh last name that much so I don't want to watch

of that but this is about the attribute of the products and how these product actually has been purchased based on uh the attribute comparison so it's about actually understanding which detail of the product is the most important as much as possible that's why if you check check gbt store you will see some of my agents one of them is about actually the product attribute generation it is able to it is for actually explaining which attribute of which product is more important this is about product classification based on internet product search and here you will see

that contextual similarity of existing parameters are important if the products have similar attributes it will be important another important thing is if you have the multiple versions of the same product as actually theam book.com for instance just for little PR we have 600 different versions which means we have actually six different attributes or attribute values which means we have really rich amount of relevance for the entity the problem is it's not Consolidated but by by adjusting your internal links you can actually uh consolidate it further in this case and here if you check the decision

trees here with these yes or no type of uh sections it will be helpful for you to understand how they are actually using product data on your pages and I suggest you to read actually detailed description through the uh through the claims especially and through the drawings as well in these areas you see directly how they actually use the product parameters to associate products to each other or classify them you this will also be helpful for you to understand what type of products your actual competitors are selling in e-commerce for instance they might be also

selling uh some some let's say pencils or some book reading created materials not just the books but book reading materials or certain type of Ni say University exam studies for instance if they if the specific user cluster searches for this these things actually expanding your product coverage also will be helpful for you to rank higher it's like topical coverage from the e-commerce point of view and this one is about product variance similar to the previous ones actually that's why sometimes you will see that similar people or uh similar departments are appearing in this area this

is actually mainly about finding the same product but it's different variations and once we have different variations of the same product you will realize that the search engine classify all these web sources and they are trying to understand actually product and the product query relevance between them and how they can actually uh compare these sources to each other you don't have to actually read everything here but even if you are able to read this section it will be helpful for you to see what type of metals they can actually be using in these systems and

the last one is about product attribute selection so if you look at these sections you will see that they are connected to each other identifying the product variants and again autom product attrib selection once you identify the product then actually once you also understand the contextual similarity between them and uh parameter similarity between them then rest is actually understanding which which parameters are more important than others plus it's also about a little bit the probabilities if you read here probability distribution for category using rega Expressions so if you look at the product attributes you will

try to understand what type of other group of products can be similar to that and also what type of e-commerce Pages probably it they will be satisfying that even further and further if you go to the for instance uh let's say here and if you do a simple search let's say organic baby cream and in the United States in this section you will directly see that they are getting these parameters and these parameters are actually coming from all the ranking documents as much as possible and if you look at the results they directly actually put

Amazon over other sources and they are not doing this because they like Amazon definitely they do it first of all there are enough level of user generated content secondly there is enough level of sales there and thirdly and most importantly Amazon has nearly all the variations or the variance of these specific products then after the Amazon we can see maybe Target or Target and more and the other results are actually also uh connected to that angle to for instance again here the best context and enough level of reviews and a kinds of a price directly

coming here together with the shipment information as well and if I write here a number let's say nine organic baby creams or something like that in this case is I search from this area the results will be slightly different but once we have the ACC of number this source is directly coming here because the name is the is has a number like eight and it's Clos number to the nine then we directly see that number eight again close to the first one because once you choose the first one the second one will be coming based

on similar to the first one then we again we have this 91 and again this time we have this specific Source tiger baby nutrition baby cream for body Etc the important thing here is that according to the number that you have chosen here the similar similar ones are coming here but according to the first one the type of the content or document also will be changing this one is similar to this and this one is similar to this and also closer to the query because this is mainly about a kinds of comparison and kinds of

let's say a simple affilate type of listicle this is also similar to the that one too then we have again the Amazon but a kinds of let's say individual product page and if you open that again you will see most of other similar products that directly appears in this area or in other sections in Amazon as well so whenever you have more products in your website or whenever you are able to cover most of these attributes or this combination of attributes and parameters you will be able to actually rank higher too when it comes to

the e-commerce whether for health whether for a different industry I strongly suggest you to check this case is also in the Bel at as your case study list you can check it if You' like to join our community check our topical Authority course you can join our community and join also the weekly meetings that our community is performing and if you want to join to the public Community subscribe to our newsletter I thankful that actually we are growing right now we are growing like over 3 to 4,000 members across different platforms and over 1500 of

these members are in the premium Community as well with that said in the next video I will be sharing some of the successes from our community members as well and then I believe or maybe during this video even if I am taking it like way earlier probably I will be in changm so I hope we will meeting in our let's say meet up in chm as well thank you for watching and see you in the next time