Dan Russell Lecture

Google search guru Daniel Russell spoke on the The Revolution in Asking and Answering Questions in this video – showing how new search capabilities empower us to ask new questions and get “deep answers”. Video lecture has many good examples from Google search. Closes with demo of conversational style for question and answer.

The Revolution in Asking and Answering Questions, Search Research (May 13, 2014)

Google’s Fact Extraction

Bill Slawski (SEObytheSEA) examines fact extraction process in Google and notes a way that Google may be using  anchor text to find other terms for those facts.

  1. HOW GOOGLE MIGHT IDENTIFY SYNONYMS FOR ENTITIES USING ANCHOR TEXT, (June 4)
  2. FINDING ENTITY NAMES IN GOOGLE’S KNOWLEDGE GRAPH (June 5)

He explains that there are two types of crawling and indexing.

  1. Indexing text – as we are accustomed to thinking – and ranking pages.
  2. Identifying named entities – people, places, things – and picking these out as “fact extraction”. These crawlers have been called “janitors”.

Google is likely using anchor text to learn more about the entities and specifically to gather alternate names.

Entity-based search

SEO analysts are being encouraged to design for entity search – and searchers can keep in mind many of the same points.

Feeding the Hummingbird: Structured Markup Isn’t the Only Way to Talk to Google by Cyrus-Shepard, The Moz Blog (June 17)

Structured markup   does help the search engine   identify the entities: people, place, products, events; but the search engine can parse text to identify subject – predicate – object in the query and in the text. This is a triple – and Google and others are good at extracting the semantic meaning. The advice to the SEO analyst to include “appropriate predicates and objects” might apply to the searcher too.

Meaning is clearer with use of synonyms and context – the “co-occuring phrases”.   Google excels at expanding a word in context to related words – synonyms.  The SEO analyst and the searcher could help the process by adding a few too.

 

 

Bing at Five Years

Bing, though it has not upset Google in market share, does have some successes in creating a search platform to be used in other applications.\

At Five Years Old, Bing Has Come Far Yet Has More To Grow, Danny Sullivan, Search Engine Land (Jun 2)

“It just scored another win by powering Spotlight search in the forthcoming versions of Apple’s Mac OS X “Yosemite” operating system and iOS 8. Bing has certainly proven itself to be more a contender than I think many would have believed. Maybe in five years, we’ll see a real platform-to-platform battle happen.”

And in another five years the search engine scene will be completely different.

Google’s penalties

Danny Sullivan examines the impact of Google’s algorithm changes on the news and community weblog, MetaFilter. MetaFilter, a long and well regarded service, has suffered from severe decline in traffic for reasons unclear to the owner. Sullivan tries to sort this out – is it a penalty for ads, for thin content, for “inorganic” links? Is this the result of a manual intervention or an automated filter? Nothing stands out (at the time of the posting) and Google has been slow to help. The article does show the enormous complexity of assessing sites and the damages that can ensue for a business.

On MetaFilter Being Penalized By Google: An Explainer, Search Engine Land (May 22)

BIng and Entity Search

Bing is driving forward with entity search. While Google wants to be an “answer engine”, Bing is saying it is a “do engine” for helping people take action – book that restaurant reservation.

Microsoft Has Big Plans For Bing’s Entity Engine, Frederic Lardinois, Techcrunch (Mar 30)

Microsoft clearly has big plans for using entities in Bing and products that rely on it; the company plans to open up a part of this entity engine so more third-party sites will be able to highlight some of their features on Bing.

Expanding the Knowledge Graph

Dr. Meyers finds that Google is much better at understanding questions and could be revolutionizing organic search.

Knowledge Graph 2.0: Now Featuring Your Knowledge, Dr Peter J Meyers, The Moz Blog (Mar 25)

But where does Google get the information – from humans!

The main sources of data for the Knowledge Graph are curated by people. Ironically, Google is facing the same dilemma with Knowledge Graph in 2014 that led to the creation of internet search engines in the first place.

But that will have to change. Meyers gives some examples of direct extraction of answers from pages.

Accentuating the entities

Barbara Starr continues her examination of the adoption by search engines of practices that make semantic web a reality. This includes “understanding” the queries through word connections, returning “answers”, and changing the display. Now – webmasters should add semantic / structured markup (using schema.org) to their pages in order to get their content attached to this ever-growing knowledge base.

How To Tell Search Engines What “Entities” Are On Your Web Pages by Barbara Starr, Search Engine Land (Mar 21)

Be sure to read the summary / takeaways at the end of the page.

We are heading toward a significant proliferation of markup on the web.  For example, on the web, the same event may be listed on multiple websites. In these cases, the markup from the most official website would be used in the knowledge panel.

Bing is very social

Bing has a strong relationship with Twitter and Facebook – enough to index “half a billion tweets and two billion Facebook updates every day” and use social signals in assessing authority of pages.

The Intersection Of Search & Social, Casie Gilette, Search Engine Land (Mar 13)

Bing indexes half a billion tweets and two billion Facebook updates every day. Bing also integrates with LinkedIn, Klout, and are looking at Google+ data. That’s a lot of information!

 

According to Forrester, Bing is using all of these social signals to help determine authority. The search engines want to provide the best content from the best people, and social can help them determine who/what that is. Forrester noted that social signals are in fact influencing rankings, but only as part of an overall pattern. For example, when a large number of people are sharing or linking to a site, this attracts Bing’s attention. To determine if it truly has value, they may try to rank it higher and see what happens. He did however say that there is not a direct connection between number of likes and tweets to rankings.