Hummingbird made semantic search possible - understanding entities & relationships.
Rankbrain, BERT, etc all help Google understand semantics of language.
MUM will take this even further.
Google is organizing content around entities.
2/20
The entities associated with your business online create a digital picture of your brand.
Your "brand" is important. When people perceive a brand, they think of its expertise, its authority, its popularity and whether it's trustworthy.
3/20
Remember the Vince update in 2009? Eric Schmidt who was CEO of Google at the time said that understanding brands was the key to keeping the internet from being a cesspool of false information.
4/20
From Olaf's article: "Google cannot afford to place content from untrustworthy sources in the user's field of vision. Especially for YMYL topics."
5/20
Back to semantic search - a semantic database is organized out of entities, their relations and attributes.
If something is recognized as an entity, Google can quickly access all the stored information about that entity and its relationships with other relevant entities.
6/20
If a search query contains an entity, Google can quickly access all stored information about that entity and its relationship to other entities in the Knowledge graph.
NLP is really good at extracting entities.
7/20
Here's a great representation of how entities and the relationships between them can be pictured in a knowledge graph.
Here is another example showing how the entities Taylor Swift and Joe Alwyn could be organized in the knowledge graph.
9/20
Olaf thinks there are more entities that exist than the ones that are officially stored in the knowledge graph.
This is interesting to me because this week I was pondering that as well. I don't come up in a kg API search, but I bet I'm still an entity in Google's algos.
This is where we start talking about the "Knowledge Vault" - much larger than the knowledge graph.
Olaf thinks (and it makes sense to me) that entities that aren't fully validated so that Google has full confidence in them, are stored in the knowledge vault.
12/20
Since Hummingbird, Google has sought to identify, extract and understand the relationship between entities.
The more frequently co-occurrences occur (i.e. this entity is associated with this topic), the greater the probability that a semantic relationship exists.
13/20
For example, if the "Empire State Building" is often mentioned together with "skyscraper", this is likely a strong relationship.
"Google can use the strength of these relationships to assess expertise and, above all, authority and incorporate them into the E-A-T concept."
14/20
Google's keyword planner can show you keywords that are associated with your brand.
If you remove your brand name from these keywords, you'll get a list of generic keywords that have semantic relevance in relation to your domain.🤯
15/20
What does all this mean for SEO?
Your brand as represented online and its authority are very important.
SEO fundamentals like crawlability, indexing, internal linking and website hygiene are important, but the triad of relevance, trust and authority matter incredibly.
16/20
This means that good PR (not pagerank, but public relations) can be powerful.
"One should try to generate cooccurrences and links from topic-related editorial environments via PR campaigns or content marketing campaigns."
17/20
Someone shared this with me today. I thought it was fitting.
Matt Cutts recommending white hat link building years ago!
Get links that are editorially given on the basis of merit. Links mean that your content is trustworthy.
Olaf ends with this: "The more clearly Google can identify the positioning of the company, author and publisher, the easier it will be to rank the thematically relevant content linked to this entity."
I had a conversation with the AI about Google's algorithms.
I asked if paid links help in SEO. It seemed confused.
🧵
Oh, Google's guidelines? Tell me more. Why should I follow them?
I asked how often it happens that a site is penalized for violating Google's guidelines. (Google penalty remediation has been my jam for 10 years now.)
There will be some questions in the thread, but this isn’t a typical Q1/A1 chat.
I don’t claim to have all the answers here, but want to generate discussion. My opinion is that SEOs who understand semantic search have a big advantage. We should not overlook it!
I’ll end the tweet thread by asking for discussion on how we can practically use semantic SEO to help our clients better.
Q5 While not mentioned in Callum’s article (will be in part 2), it’s worth discussing the product reviews update. The most recent release was March 23. Did you see sites affected? What have you seen?
A5 (cont) We also had some clients with losses. Most were the manufacturer or vendor of a product. They lost positions to content that reviewed multiple products for review queries.
A4 While not every site that is negatively impacted by a core update can recover, many can. In some cases improving EAT helps. Google has put a big emphasis on content as well.
Google’s advice: Focus on Content and E-A-T #learnseo
Q3 E-A-T is not an algorithm, but assessing your content in terms of what Google has described as E-A-T may help it rank better.
Have you had success in improving E-A-T? What did you do? #learnSEO
A3 We have had great success in helping sites improve by working on E-A-T: Adding expert authors if appropriate, cultivating good reviews, schema to improve entity understanding, mentions in authoritative places and more.