On Apr 3, 2018 Google’s head of search, John Giannandrea, stepped down and was replaced by two Google veterans; Jeff Dean who now lead Google’s AI efforts, and Ben Gomes who became became VP of search.
John Giannandrea replaced Amit Singhal who stepped down in 2016 to go to Uber. Singhal rewrote the original Google ranking algorithms, and John was the AI guy who was supposed to have driven AI into Google’s most important core product – Search.
Communication from John Giannandrea came altogether too slow if at all: Though Google the company constantly talk about AI, the messages rarely issued from John Giannandrea’s preciously brittle lips.
Google’s new head of search, Ben Gomes, on the other hand, is an absolute ledge: He has been with the Search Engine giants since 1999, one of the early Googlers, who didn’t drink his coffee with un-lactated soy bean content and sat towards the front of the table during search quality meetings.
High level: How Google search works
Someone typing “volleyball movie on an island” and getting the result, “Cast Away” thinks Google understands what they meant. It doesn’t.
The search engine has no idea what those words mean. Google looks for pages where those words their synonyms, or even common misspellings appear, and surfacing what they deem the most relevant according to a list of over 200 signals. Those who try and manipulate these signals to ensure a webpage ranks at the top of Google for commercially relevant queries are known as SEOs – an acronym for someone in the SEO (Search Engine Optimisation) industry.
Google bots, known as web crawlers, cover the internet at a rapid rate of nots, gathering chunks of information from hundreds of billions of webpages. Google then stores that data in a constantly changing index, which is a list of URLs, taking note of signals like freshness and how often that URL is cited. When you enter queries into Google’s search bar, the query string is fed through a set of rules and processes known as Google’s search “algorithm.”
This process compares your query against the data in Google’s index and the results most likely to satisfy your query are returened to you in order of relavancy. The algorithm decides the order in which documents are served. All thgis happens in a fraction of a second.
This algorithmic process has been becoming more often about Google dominating its own homepage with ‘query refinement links’ that take a user through to subsequent Google searches or other Google products. The information served here by Google is scraped content from other large websites, such as Wikipedia, but also from smaller resources (like this website). Iterations to Google’s algorithm that changes how it serves content occurs ‘more than 2000 times a year’ according to their Search Liason, Danny Sullivan, and are known in the industry as ‘broad algorithm updates’.
You can read about some of the major algorithm updates of 2018 here.
Two important changes to the way Google works which facilitates their dominance of online attention
2007’s Universal Search, which wove elements such as links, news, images, and videos into one list, ended the era when Google results were just “ten blue links” and was a huge advance for Google’s domination of the internet. It meant users were served a variety of media to help answer their query, rather than making them perform subsequent searches or using alternative search engines. The second change, 2012’s knowledge graph, allowed Google search to directly answer questions such as “What’s the population of Monaco?” which helps confine users to Google’s walled garden of scraped content.