How does Google work
- 1. How does Google work? O.o
- 2. Goal Provide a fundamental understanding of Google search
- 3. What does Google want? To be the world’s best search engine by… ◎ Providing the useful and relevant results to their users ◎ Giving higher rankings to high quality sites that people are interested in ◎ Getting all of your personal information and tricking you into doing things
- 4. 60 trillion Individual pages that make up the interwebs
- 5. Back in the deezy Crawl the web (30 days) Index content (2- 3 weeks) Check for updates (PageRank) Produce results from index
- 7. When Google crawls ◎ Web content is indexed based on keywords ◎ The index notes things like word order, proximity, frequency, association with the page title/heading, etc. ◎ When you run a search, you are searching Google’s index
- 8. When you search Algorithms ◎ Attempt to decipher the intention behind your search ◎ Perform document selection based on their interpretation of your query ◎ Selected documents are sorted and pared down based on their determined relevance Results ◎ Are run through ranking factors like • Quality • Freshness • Security • Context (i.e. region and personal history) ◎ 200+ factors
- 9. Place your screenshot here SERP Displays results like snippets, articles, immediate answers, graphs, images, ads, etc.
- 10. End You know stuff now
Editor's Notes
- #4: This slide aims to provide context for the explanation of Google’s process by describing their goals.
- #5: Just a fun fact, but also mentioning that this is the amount of content that Google is attempting to catalog.
- #6: A visual representation of how Google worked when it first started: Their bots crawled the web over 30 days After being crawled, it would take weeks to index the content In that time some pages would have been updated, so they would go back and check for updates prioritizing sites with high PageRank Search results would be pulled from that potentially outdated index
- #7: This is a view of Google’s order of operations. Many people are under the misconception that Google searches the entire web on the fly and spits results back to you. However, in actuality, crawling means that Google is placing pages into its index and that’s where your search is actually run. Ranked results are produced on the fly, but they are pulled from the index where parts of Google’s algorithm have already been applied.
- #8: Pages are sorted within the index by content type and a number of other factors, making it easier to retrieve relevant documents when a search is run.
- #9: When a search is run, Google pulls all potentially relevant documents and then narrows down and ranks the results based on perceived relevance to the user’s particular search. 200+ ranking factors are used (Google at one point has said there are around 10,000 ranking factors).
- #10: This is an example search for the phrase “kate spade shoes.” When pulling documents from the index, Google might initially look at any pages that contain the word “kate,” “spade,” or “shoes.” It would then pull common pages out of those three sorted categories (pages that contain more than one of the words). Then it sorts by algorithmic application as well as looking at the word order and perceived intent of the search.