Hey Dan, have you discovered a tool replacement for this yet? I would love to explore implementing entities into my content development and other SEO processes, but without a tool I'm not sure where to get started. Thanks for great/informative content as always!
Thanks a lot! There's lots of videos exploring this topic but I think you killed it here. As you mentioned with the repeated entities - there's still so much I want to learn.
Although there are now tools that can do this for you in a structured manner, this is STILL valid. Great video Dan, wish I'd stumpled upon this back when it was published.
Thanks! I think some people had honest concerns that I was neglecting to clarify the "real" way to use the tool/data/API is to analyze large sets of content etc -- however, I do think some people just assumed I was calling these things "ranking factors" etc or didn't really watch the video and assumed I was making invalid claims. But to this day I still use these methods intelligently to help inform ways to improve content!
i'm curious about how much Google uses this particular product in their Search Algo. For example, try searching "uncertain weather" on Google and you'll see a serp feature and organic results from weather related websites for a place named "Uncertain, Texas". However if you put "uncertain weather" into their tool, cloud.google.com/natural-language/ it won't detect a location. Try this with more recognizable names like Seattle with "seattle weather" and the natural language tool notices a location. I'm not personally in Texas and I've never been to "uncertain" so it seems like Google is determining intent exclusively from my search phrase: "uncertain, texas". If it was only for the weather serp feature, I might say they have a separate algorithm here, but since the organic search results show local weather information to Uncertain, Texas, I have to assume there's something there too. If I had to guess, it shows that either 1) the Cloud Natural Language product is a separate system from what they use internally for Google search or 2) Google still falls back to pattern matching / keyword matching in many cases and only uses entity recognition as a secondary factor to determine "what something is about". Another example to think about: Search (A) "name of the main pointy ear character from star trek" Compared to a search for (B) "spock" conceptually these are searches for the same entity, but google's organic results show you more pages that talk about the ears in search A as opposed to B. What do you think is going on in these examples?
Hey Hayk - this is a really good point and observation! If I had to quickly guess, for "uncertain weather" (really interesting example!) they are probably leaning on a lot of user metrics (people way back may have searched "uncertain weather", then refined it to "uncertain weather in texas" etc) and the more user behavior maps to wanting the weather in a town called "uncertain" google eventually figures out a strong intent. It's also a good question if Google uses the NL tool in their actual algo. I think understanding a topic is a bit different than putting ranking scores on a document. ie: links give weight (and some context too of course) but understanding language does not assign weight, rather just improves understanding to which then ranking weights are still applied on. So if Google is using this exact tool or not, I certainly find it best to focus on picking apart the feedback it provides tot analyze my content and competitors etc to then make content better.
Dan, dude...THANK YOU. I had heard about this but hadn't explored... I'll take this over gold any day. There is a massive mega-trend going on with Ggl and other SEs, diversifying away from links and re-weighting content and on-page. The pace of this change may be slowed by (their) current technology, but there is little doubt of where we are headed. --Question: on the first (entity) step, when you pulled the page content from Wikipedia, did you include the table of contents list or just the body content? --Another question: does Ggl offer an API for this and are any of the tool providers incorporating? I only know of Keyword Keg as a vendor that incorporates on-page (into their KW difficulty scores) but it seems they are only doing simple scraping and ratio analysis. I think people are in denial about where things are going. There's a generation of "experts" and a whole industry around the importance of links. No doubt, they will be a factor for a long time yet. But I think the lack of good tooling around quantifying on-page quality makes it even harder to prepare for this future (for those ready to embrace it).
Hello! 1) I did not include the table of contents, but I could see choosing to include that or not as a good strategy for looking for entities. 2) There is indeed an API, but I have not tried it yet. All the info is here cloud.google.com/natural-language/ I agree for sure that on-page analysis is really important, and I'd love to see more people using it and talking about it. I think that what a lot of people call EAT is really better understood by search engines by analyzing content (vs looking at things like links, author pages etc).
Howdy - I am aware this free tester has been removed from Google's page. If I find any good replacements for it I will post it here.
Hey Dan, have you discovered a tool replacement for this yet? I would love to explore implementing entities into my content development and other SEO processes, but without a tool I'm not sure where to get started. Thanks for great/informative content as always!
@@tylerrouwhorst11 The free tester should be active again, Google put it back a few weeks later!
@@evolvingSEO I see that now! Thank you!
@@tylerrouwhorst11 Where is the tool? I can't find it.
Thanks a lot! There's lots of videos exploring this topic but I think you killed it here. As you mentioned with the repeated entities - there's still so much I want to learn.
Although there are now tools that can do this for you in a structured manner, this is STILL valid. Great video Dan, wish I'd stumpled upon this back when it was published.
Appreciate that!
Fantastic. People who dislike your video are just haters.
Thanks! I think some people had honest concerns that I was neglecting to clarify the "real" way to use the tool/data/API is to analyze large sets of content etc -- however, I do think some people just assumed I was calling these things "ranking factors" etc or didn't really watch the video and assumed I was making invalid claims. But to this day I still use these methods intelligently to help inform ways to improve content!
The Google NLP demo tool is gone. Is there any service or a user friendly script for purchase that pulls the entities the same way?
I am not aware of one at the moment! If I find something I'll post a comment update here.
@@evolvingSEO Thanks, please do! It would most likely be a pretty difficult DIY trying to get sth working just by their API documentation.
And is not ranking No1 because it has great backlinks (ha ha)!
I'll count you!
This is awesome! Thank you so much.
Awesome - you bet!
i'm curious about how much Google uses this particular product in their Search Algo. For example, try searching "uncertain weather" on Google and you'll see a serp feature and organic results from weather related websites for a place named "Uncertain, Texas". However if you put "uncertain weather" into their tool, cloud.google.com/natural-language/ it won't detect a location.
Try this with more recognizable names like Seattle with "seattle weather" and the natural language tool notices a location.
I'm not personally in Texas and I've never been to "uncertain" so it seems like Google is determining intent exclusively from my search phrase: "uncertain, texas".
If it was only for the weather serp feature, I might say they have a separate algorithm here, but since the organic search results show local weather information to Uncertain, Texas, I have to assume there's something there too.
If I had to guess, it shows that either
1) the Cloud Natural Language product is a separate system from what they use internally for Google search
or
2) Google still falls back to pattern matching / keyword matching in many cases and only uses entity recognition as a secondary factor to determine "what something is about".
Another example to think about:
Search (A) "name of the main pointy ear character from star trek"
Compared to a search for (B) "spock"
conceptually these are searches for the same entity, but google's organic results show you more pages that talk about the ears in search A as opposed to B.
What do you think is going on in these examples?
Hey Hayk - this is a really good point and observation! If I had to quickly guess, for "uncertain weather" (really interesting example!) they are probably leaning on a lot of user metrics (people way back may have searched "uncertain weather", then refined it to "uncertain weather in texas" etc) and the more user behavior maps to wanting the weather in a town called "uncertain" google eventually figures out a strong intent.
It's also a good question if Google uses the NL tool in their actual algo. I think understanding a topic is a bit different than putting ranking scores on a document. ie: links give weight (and some context too of course) but understanding language does not assign weight, rather just improves understanding to which then ranking weights are still applied on.
So if Google is using this exact tool or not, I certainly find it best to focus on picking apart the feedback it provides tot analyze my content and competitors etc to then make content better.
Great video Dan! Very interesting.
Thanks Mike!
Dan, dude...THANK YOU. I had heard about this but hadn't explored... I'll take this over gold any day. There is a massive mega-trend going on with Ggl and other SEs, diversifying away from links and re-weighting content and on-page. The pace of this change may be slowed by (their) current technology, but there is little doubt of where we are headed.
--Question: on the first (entity) step, when you pulled the page content from Wikipedia, did you include the table of contents list or just the body content?
--Another question: does Ggl offer an API for this and are any of the tool providers incorporating? I only know of Keyword Keg as a vendor that incorporates on-page (into their KW difficulty scores) but it seems they are only doing simple scraping and ratio analysis.
I think people are in denial about where things are going. There's a generation of "experts" and a whole industry around the importance of links. No doubt, they will be a factor for a long time yet. But I think the lack of good tooling around quantifying on-page quality makes it even harder to prepare for this future (for those ready to embrace it).
Hello! 1) I did not include the table of contents, but I could see choosing to include that or not as a good strategy for looking for entities. 2) There is indeed an API, but I have not tried it yet. All the info is here cloud.google.com/natural-language/
I agree for sure that on-page analysis is really important, and I'd love to see more people using it and talking about it. I think that what a lot of people call EAT is really better understood by search engines by analyzing content (vs looking at things like links, author pages etc).
Sir how can I copy the analyzed text and trees to .doc file so I can keep it for offline use ?
thx bro
What the color shows?