Google Algorithm Leak: What the API Document Reveals About SEO

who would have thought that the most prized algorithm in the world would be accidentally leaked to us via a GitHub depository and that that leak confirmed Google and its engineering team have not only been making false statements to us for a very long time but that a lot of what we as an SEO Community have been showing as Theory or research was actually reality all along so in this video I'm going to be breaking down the

facts we have so far what this means for us as SEO professionals as as a whole and a bunch of my own research and insights into the specific ranking attributes that we can learn the most from this leak so without any further Ado let's jump into [Music] it so what do we actually know so far well this wasn't a direct leak in the sense that somebody internally leaked it out on purpose it was accidentally distributed via

a GitHub depository and hilariously it was given a distribution via the Apache 2.0 distribution license which meant that it was actually given a license to anybody who downloaded it directly off of that Hub so you actually have got a right to do it and Google has no right to kind of ask for that data back it's just publicly in the realm now the data is at maximum 9 months old because it references parts to August last

year there's not actually 14,000 attributes like everybody has been kind of quoting around the SEO Community because not all of them are search related only around 7,300 of those attributes are actually search related and then the other attributes are for the various other Google ecosystem like maps and YouTube so as SEO professionals what does this mean for us as an industry well well actually it reinforces a lot of what we already knew like I already said

a lot of the attributes that we are seeing directly correlate with a lot of the research and real world Ser analysis and data that the community has already been sharing for years and years at this point and it just reinforces a lot of the findings that we have already found from our own research it also gives us more insight into what Google is currently targeting to try and demote in its rankings especially how Google deals with

ankitects and spam which we'll get on into a little bit later on and we can also take a look at a lot of these signals as being in production due to comment like this where it says that this specific attribute is not in production and it won't be populated live that means that we can kind of take into consideration with a slight pinch of salt that most of the other attributes are actually live and in production

on the algorithm and bear in mind that some attributes do contradict each other so you just want to make sure that you are you are kind of in the know against all of these different attributes and that you are taking into consideration that some attributes May contradict other ones and they may affect other ones they may be more important and have a higher scoring system than others because we don't currently have access to any sort of

scoring data or scoring system in relation to these attributes which is kind of the real secret Source behind a lot of this though this attribute leak is massively important to the SEO community and just reaffirms a lot of what we have already known for a long time so one thing I want to talk about before we actually get on to some of these attributes themselves is that actually a lot of the attributes that you're looking at

individually might not make all that much sense or could be misinterpreted as something else but if you look at the surrounding attributes it makes significantly more sense what that attribute is specifically being used for and what it is already been tied into one of the benefits that we have from this leak is that it's all modulized so we can actually go in and see what the reference title is for where that attribute is within a specific

module and kind of reference that so this specific one where I'm referencing here is within the anchortext module so now we can have a look at the actual attributes themselves the very first thing we can say and a pretty obvious one that we all knew for a long time though Google's Engineers themselves said that it didn't exist site Authority domain Authority and the fact that you have it and you get naturally and organically promoted just because

your website has an authoritative link profile it exists it's real we can confirm this one we can take it off the Box something that was a lot more hotly contested was the fact that toxic links were real or not and if a back link could actually have a negative effect it is confirmed that at least in one instance that there is some form of a penalty applied to a bad backlink scenario and if you look just

under that specific attribute you'll also see that there's a page level penguin penalty which again is something that I've talked about for a long long time where we have found that P penguin and especially rolling penguin was a page level penalty and that's why we would always recommend that when you building links that you would build them at supporting content and not your money Pages because it wouldn't affect the money Pages if via internal linking it

would only affect the supporting content because it's a page level link penalty but this just again reaffirms toxic links are real and you should probably disavow links just as toxic links are real so are protective pillow links pillow links seem to have some level of protection especially if it says and I'm quoting here the dock is protected by goodness quote unquote of early anchors that essentially to me says two things that it is reinforced Again by

other attributes that talk about trust signals and that you want to have trusted sources but also that if you are using a lot of those kind of branded and more white hates anchors especially earlier on in your campaign that you can tend to get away with a bit more higher levels of toxicity higher levels of exact match higher levels of P partial match further on down the line and that's exactly why we were using these pillar

links throughout our SEO campaigns another one that was very hotly contested and that had some serious debates throughout the SEO industry's lifespan was the Google sandbox and if you actually go and Google Charles float or your YouTube search Charles float sandbox you'll see that I have videos to about how to try and at least beat the sandbox and especially when you are looking at repurposing age domains or expired domains you are much much more likely to

get this kind of uh sandbox affecting you so in general it seems to be towards newer domain names and AKA fresh spam and making sure that it doesn't deliver new websites that could be potentially harmful to users or harmful in terms of spam or harmful in terms of hacking and things like that but in general the box exists and it's something that the Google engineers and John Mueller specifically denied repetitively and again Google claimed that page

rank was dead but according to these leaked documents it seems that actually page rank or the old version of page drank was deprecated in replacement was a newer version of pay drank which took into account this nearest seeds protocol which for those that don't know seed sites are these original websites that Google class specifies to start crawling from the search engine if you imagine if you're just starting a search engine you need to put websites into

it for it to start crawling put URLs in for it to start crawling and Google will put seed sites in there for it to start crawling and it puts weight on those seed sites in general you can think of seed sites like the New York Times the BBC and Wikipedia in general the higher you the more links that you have from those websites and the closer your links are to those websites so the more links they

have from those websites to your back links themselves the better your links are the better your quality link profile is and the more trustworthy it is the safer it is to rank etc etc etc probably one of the most interesting parts of the doc actually was this comment on freshness being a ranking signal that isn't particularly sensitive I guess that Google does actually classify specific types of signals and specific types of ranking attributes as more sensitive

for example it's it denied for the longest time that click data was a ranking signal and we will get onto that in a little SEC but the doj trial kind of forced Google's in force Google's hand into admitting that click data was a thing it seems here again that freshness is a thing but it's not a particularly sensitive thing and there's definitely been a couple comments from Google Engineers over the years that freshness has been taken

into account and that it will specifically help boost around kind of current events or any kind of trending topics another one that I've been screaming from the rooftops for years but trust signals are real and that they can definitely help protect your website and kind of future proof it against potential demotions or other kind of spam capabilities that Google is looking at trying to Target sites out there with this one specifically is referencing anchit texts more

so than General kind of trusted links but in general you can see at the bottom there it says total number of trusted sources for this URL so in general you can see that trust signals are a massive part of how Google looks at dealing with Spam and especially if you have a new website or especially if you're in ym niches you're wanting to be taken advantage of these kind of trust signals as much as possible and

by all means it looks like Google just wants you to build as many of them as possible which is what I've been saying around entity stacking for a while now and like I said earlier Google definitely uses click data it definitely uses Chrome data it's a pretty obvious sign that Google kind of created the world's biggest browser and if they weren't to use that data especially the fact that they are allowed to use that data it

wouldn't really make much sense yes their engineers and their kind of PR team their Comm team have been denying it for a very long time but to a lot of people it was pretty obvious and it was one of the main reasons why CTR manipulation was such a problem for a very long time and it's kind of still a problem when it comes to negative SEO even to today however in general what we can see from

this is that Google has got a lot of different factors when it comes into this specific case and that some countries are prioritized other over others some clicks are inherently bad some clicks are inherently good and it kind of takes into consideration as well the last longest clicks which was the number of clicks that were last and longest in related user C AKA how many times the user actually lasted on your page versus other pages and

it does seem to take into consideration some sort of correlative effect as well so clicks data definitely seems to be a big part of the algorithm and like I said earlier the doj case did expose that we just did know to kind of what sort of an extent but nav boost which is the system that does take into consideration this click and Chrome data definitely seems to be one of the bigger systems inside of this league

so we do actually see that Google definitely does censor content as well it does seem to be the case that it might be more current situation related so things to do with elections things to do with covid but there also does seem to be some sort of a universal whitel list just allowing websites by default to be entered into every kind of query and making sure that they aren't pruned out of those queries by mistake however

I'm not sure how legal this fully is because Google as a search engine is supposed to be independent though they are supposed to have some kind of level of cleanup when it comes to anti- Illegal content but I'm not sure how much necessarily things like covid and election misinformation is actually illegal and that might actually be a bit of a gray area when it comes comes to Google's ability to actually censor it search engine like it's

doing in this specific case so a few people have been calling this one the Neo Patel score or the copycat score is its actual name but it seems to be the case that Google is actively trying to drank duplicate content drank kind of syndicated content and it would make sense however there is still linked version to take into consideration where actually it doesn't necessarily matter who was the first person to publish the content and who is

the copycat it's more so who is who has the strongest link Authority especially at a page level and that tends to take over from this score especially from our own testing however it's pretty interesting to know that duplicate content definitely does get dranked at some level you can also see that ugc signals got absolutely turbocharged and there seems to even be a comment around it being times 1,000 and flawed you know you can interpret what that

means by yourself however my own interpretation kind of leads me to think that there is some flawed nature in their ugc page quality signals and that they have increased it and turbocharged it to the absolute Moon and that would show on why a lot of these Cups have been taken over by the likes of Reddit and core and why Reddit it self has seen upwards of a 600 to maybe I think 800% increase in organic traffic

over the last year one of the parts that was very very interesting to me as well which again kind of reinforces my entity stacking protocol for every new website and every website in general that's trying to rank in Google is this document that shows how Google actually build document or the NLP Saft entity document it shows all of the various attributes around how Google is correlating and entity including things like gender mentions profile references and representative

mentions and Salient so the the scoring of how relative it is to The Entity score as well and all sorts of additional ATT tribut as well but it's been quite common knowledge now for a long time that Google has been trying to build the world's biggest entity base and that that will help power the AI uh for future that Google is trying to build and it will be one of the biggest resources that Google has at

its disposal and that I think it has definitely achieved it so far with the algorithm but it's just interesting to see the specific types of uh features and attributes that Google is looking for when it's building the entity so it helps us further reinfor course how we are approaching building entities in Google search and then finally I want to take a look at how Google itself looks at anex because we got some really really interesting information

about how Google treats ankitects and that anex is actually one of the most important parts of SEO and your link building uh and your link profile and that you want to make sure that you are following these rules as much as possible and that that does mean anext in relation to both link building and internal linking but the very first one want to talk about is this creation date and actually further on down the line You'll

see deletion date and there's another one that's called Fresh date as well and that basically is showing us that Google is tracking anex changes and that it is actively looking at anex changes for spam signals so that's one of the things that I was saying previously is that you don't want to actively try and change a lot of the ankitects that you've already built especially if you're doing that in bulk because that can be a a

active spam signal and an active negative signal in and of itself then you'll have this is local one which again just indicates whether the source is an internal link or an external link then context is very very interesting and there's actually two more other attributes in the documentation that talk about uh text to the left and text to the right which further reinforced that context narrative or the surrounding context of the architex and we've always talked

about how important that is and how important you can actually um Channel some of the partial and exact match anext through the surrounding text rather than making the anex itself a partial or exact match anchor and then finally it says the dropped homepage and dropped local anchor text account which is very interesting because it actually tells is that Google is actively dropping or removing a lot of internal anchors and internal um keyword signals when you're doing

internal linking which would make sense especially the homepage because it doesn't necessarily want the homepage to rank for specific keyword signals because most of the time you want to rank the homepage for brand branded searches and branded queries alone so where do we go from here then well first of all this is likely just the beginning we've only had a few days maybe about just about a week now to actually go through this document most people

are still using AI to analyze lot of the documents and I'm going to link in the description down below to a blog post from Christopher in my team who will show you how to use llms to actually analyze this this uh leak for you and how you can use tools to do it for you you don't have to use kind of these pre-made tools from the S industry you can just use chat PT um but again

we don't have any direct scoring systems so that means that we can't really attribute any direct scores or importance to specific queries or specific attributes sorry and we need to make sure that we're taking everything with a bit more of a pinch of salt but we can try to infer the scoring system and we can take the testing data and the experiments and the years and years of research that we've done as an SEO community and

just reaffirm that using these attributes but we don't want to this leak give you confirmation bias so we want to make sure that we're testing to confirm all of our hypothesises all of our theories we want to make sure that we are confirming everything via testing and that we're date to backed and that we're not just relying on opinion because it's not fact and that we need to make sure that we're confirming everything through testing and

through the Surs and bear in mind that Google have been eerily quiet and it's very very possible that a major update could come around the corner purely to use a tribute that that aren't in these documentation and make the SEO Community scramble to go through the testing phase and research phase all over again and it could be like you know going back to 2005 or the or the Panda and Penguin days all over again trying to

figure out what Google has changed what Google is targeted great fun for us not so much fun for most of the search most of the people who rely on Google as a search engine and as a traffic source so I want to thank you for watching this video I hope you enjoyed it I hope you got a lot out of it if you have any additional insights into this leak drop them in the comments down below

make sure you like this video you can follow me on X for more daily SEO insights it's _ SEO you can connect with me on LinkedIn it's just my name Charles float and if you want any of my free and premium training you can go to Charles training.com including that AI llm blog post on how to analyze this leak for you I'll see you in the next video peace