After Pagerank, Google overlayed this algorithm with the idea of the Reasonable Surfer. This presentation at Pubcon Vegas is all about Topics, links and the reasonable surfer.
2. Today’s
Agenda
25 minutes
Look at a proxy logic for a “Reasonable Surfer” Algorithm
Look at why the outbound links on your page (to other pages on
your site) can have HUGE implications for Semantic SEO and
Topics
Why Topics are replacing keywords in search
Use two new technologies that didn’t exist last year
@Dixon_Jones
3. LastYear I left
you with the
PageRank
Math…
talk@dixonjones.com
Subject: “Pagerank”
@Dixon_Jones
5. Nobody is
paying enough
attention to
outgoing links
I ask myself:
“How might a computer change the value of each
outbound link on a page?What can it measure?”
@Dixon_Jones
11. • Up Top 40 Segments
per page
• The Green line is
YOUR link
• Internal Vs. External
• TF & CF
• Inbound Outbound
Totals
• Primary Topic
• Language
@Dixon_Jones
20. What have we
found so far?
• Links out of a page all carry different
weights
• Machines cannot easily learn by
copying visual cues
• More logical to use Structural cues
which scale better (Majestic LD is a
proxy for this idea)
• Anatomy of aWikipedia Article shows
outlinks are massive
@Dixon_Jones
22. WARNING!
Wikipedia’s internal outlinks are based on Entities. Do not link
internally based on Keywords.
This is a hard concept to
get one’s head around.
Even harder to manage
on large websites.
But Fear not…
@Dixon_Jones
33. What have we
found in Part 2
• Internally linking ideas (not keywords)
can be a huge benefit for Google’s
Knowledge Graph
• Wikipedia suffers heavily from Human
Bias, so Google’s KG does too
• Tightly themed, content rich sites are
easier to build Knowledge Graphs for
• Google have a NLP checking tool
• Scaling internal links is something
Wikipedia does well… you should too!
@Dixon_Jones
34. Further
Reading
Balog, K., 2018. Entity-oriented search. SpringerOpen.
Cai, D., He, X.,Wen, J.R. and Ma, W.Y., 2004, July. Block-level link analysis. In
Proceedings of the 27th annual international ACMSIGIR conference on Research and
development in information retrieval (pp. 440-447). ACM.
Google’s Reasonable Surfer revisited. Bill Slawski. SEO by the Sea.
http://www.seobythesea.com/2016/04/googles-reasonable-surfer-patent-
updated/
Link Context and Link Density white paper (Majestic Blog)
https://blog.majestic.com/development/link-context/
To get slides: Email talk@dixonjones.com
with “Reasonable Surfer” in the title.
@Dixon_Jones
Notas del editor
25 Minutes talking. Carrie is talking about GMB a lot I think.
Some things we took away:
Really hard to predict as a human the page with the biggest Pagerank.
If you build a new site and only used Domain Authority to create links, you could EASILY have got linked from the worst page possible, even if it was from the best domain, because of the INTERNAL LINKS of the other web pages! How on earth are you going to be able to see the strength of a link if that strength depends on the internal links on an entirely different website?!
But the Pagerank math assumes every outboiund link on a page is valued equally to other links on that page.
As humans we know it isn’t true, Google alos knows.
Google has been much LESS open about its Reasonable Surfer algorithms than it was about Pagrank.
How can a machine like Google derive meaning from what appears so random? How might Google be turning all this into machine readable language at scale?
Here is one way which Majestic has tried to do this. Before explaining what we are seeing here, I probably need to explain these vizualizations…
How can a machine like Google derive meaning from what appears so random? How might Google be turning all this into machine readable language at scale?
Here is one way which Majestic has tried to do this. Before explaining what we are seeing here, I probably need to explain these vizualizations…
One of the reasons that Google pays so much attention to Wikipedia is that its structure is SO consistent. (There are a lot of other reasons, but this one is important).
Google suggested that Freebase was now Closed, use Wikidata. Wikidata is the structural backbone behind all of the Wikimedia foundation sites.
Krisztian Balog has an excellent chapter in his Free to download book on Wkipedia’s page anatomy.
27,000 words
TIGHTLY THEMED
All the content on your site talks to and around the subject.
There are headline instances of Entities (target pages)
… and Edge cases (source and sections on source pages)