Fabrice Canel - Advanced Search Summit Napa 2019
•Download as PPTX, PDF•
1 like•150 views
This document discusses the evolution of AI in search engines from 2009 to the present. It covers how search has transitioned from keyword-based to natural language to context-based search powered by AI and deep learning. It also summarizes some of the challenges in crawling the web at scale and keeping search indexes up to date. The key takeaway is that webmasters can now enable real-time indexing by using tools like Bing Webmaster APIs to notify search engines instantly about new or updated content on their sites.
Report
Share
Fabrice Canel - Advanced Search Summit Napa 2019
- 1. fabricecanel @facan The Evolution of AI in Search
- 2. fabricecanel @facan About me 22 years search veteran at Microsoft Principal Program Manager leading the Bing Web Data platform team. The team discovering, selecting, crawling, processing billions of new or updated web pages every day. Prior to Bing and MSN Search, Lead Program Manager for microsoft.com search in the very early days of Search and SEO.
- 3. fabricecanel @facan The quest for Intelligent Search
- 4. fabricecanel @facan 2009 DESTINATION EXPERIENCE The Evolution of Bing 2012 PLATFORM 2017 AI PLATFORM
- 5. fabricecanel @facan The Evolution of Search Keyword based Search Natural Language Search Voice, Vision Context-Based Search Traditional IR AUTB streams Inverted index User Engagement User clicks, metawords Inverted index AI Deep Learning Vectors
- 9. fabricecanel @facan Visual Search • Image similarity match • Deep Learning vector encoding • Approximate nearest neighborhood search • Search in several hundred million images in a few milliseconds • Significant improvement to image recall
- 11. fabricecanel @facan World knowledge Work knowledge User knowledge Microsoft Graph – Creating a more personal graph
- 14. fabricecanel @facan What is the goal of the Search Engines? Discover and Index Everything Create “the” Knowledge Graph Super fast access to information
- 15. fabricecanel @facan No. Not quite: Provide searchers with that they can that provides insights about the search queries.
- 16. fabricecanel @facan Bing and Verizon Media Bing Ads will be the exclusive platform to manage search ads across Bing and Verizon Media networks on all devices. The Microsoft Audience network have now access to Verizon Media O&O properties, available via Search workflow campaigns. Properties include Aol.com, Yahoo.com, Aol Mail, Yahoo! Mail, Huffington Post and more.
- 17. fabricecanel @facan is bigger than You think 21% DE 19% UK 36% US 18% FR
- 18. fabricecanel @facan Is AI the solution to all my Bing team challenges?
- 19. fabricecanel @facan My team mission Build for Microsoft the world's freshest, richest, cleanest, most comprehensive and most intelligent model of the web.
- 20. fabricecanel @facan Bing Content goals Webmasters SEO goals Maximize Content Comprehensiveness Have all my sites relevant URLs indexed in Search Engines Maximize Content Freshness Have all my sites newest URLs and newest updated content indexed in Search Engines Maximize Content Intelligence Have search engines able to extract the most useful information of each web page to ease their abilities to rank my content at the top of search results Your SEO goals align with my team goals
- 21. fabricecanel @facan My TOP 3 SEO guidelines table How to have your URLs DISCOVERED 1. Good links to your content 2. Avoid duplicate and useless URLs 3. XML sitemap, RSS feeds … and new abilities in upcoming slides How to have your URLs RANKED for crawling/indexing 1. Content Authority: High quality content differentiating from others 2. Content useful, sufficiently detailed and well-presented 3. Get an audience How to have your URLs CRAWLED 1. Allow us access the content (allow in robots.txt, crawl quota) 2. Guide us to the new content: RSS, sitemaps lastmod 3. Avoid craziness with JavaScript and CSS How to have your content UNDERSTOOD 1. Standard HTML is highly preferable 2. Tell us more via schema (HTML 5 tags, Json-LD, schema.org etc.) 3. Avoid craziness with JavaScript and CSS
- 22. fabricecanel @facan Web pages may appear, change or disappear on each web site at any moment Bing needs to crawl and re/crawl to keep index up-to-date. Bing needs an effective, efficient crawl scheduling policy that 1. obeys web hosts’ and Bing’s own constraints on download bandwidth 2. is efficiently computable over billions of hosts and hundreds of billions of page The challenge of Crawling
- 23. fabricecanel @facan Crawling challenges: Diversity of solutions HARD VERY HARD EASY INTERMEDIATE Web sites with great SEO • Clean URLs • RSS and sitemaps with lastmod • Static HTML Web sites with bad SEO • Duplicates, useless, junk URLs • No sitemaps, no RSS • Lot of JavaScript • Limited Bingbot bandwidth Small web sites Large web sites
- 24. fabricecanel @facan Crawling challenges: Diversity of intends SEO, Ads people “Bing please crawl a full throttle !” Web site operation “Bing please crawl a little off- peak time”
- 25. fabricecanel @facan Crawling challenges: Detect relevant content changes <p>Video provides a powerful way to help you prove your point. When you click Online Video, you can paste in the embed code for the video you want to add. You can also type a keyword to search online for the video that best fits your document. </p> <p class=ads><a href=/r/15vGdebnbf><img scr=soccerball.png/ alt=“soccer ball”/></a></p> <p>To make your document look professionally produced, Word provides header, footer, cover page, and text box designs that complement each other. For example, you can add a matching cover page, header, and sidebar. Click Insert and then choose the elements you want from the different galleries. </p>
- 26. fabricecanel @facan Crawling challenges Webmaster hints often not reliable Sitemaps are great to inform search engines about links on your site... … but a common SEO mistake is to set lastmod (last modified date) to the date and time the sitemap is generated
- 27. fabricecanel @facan Bing main crawl metrics Metric #1 Crawl effectiveness North Star: Every page in index is a fresh copy of its web version Metric #2: Crawl efficiency North Star: Crawl only updated (fresh on-page content/useful outbound links) or new URLs “Metric” #3: Obeying politeness constraints North Star: Never crawl more often than webmasters *want*
- 28. fabricecanel @facan What’s about we make it super simple! You tell us Real Time about your URLs change
- 29. fabricecanel @facan Transforming the paradigm Real/Near time indexing Day 1: content update LAST CENTURY NEW CENTURY Day 1 Day 2 Day 3 Day 4 Day 5 My daughter has won a prize at school. https://answers.yahoo.com/question/... Mar 29, 2012 · My daughter… Day 1 Hey, I changed. Thank you ! My daughter has won a prize at school. https://answers.yahoo.com/question/... Mar 29, 2012 · My daughter…
- 30. fabricecanel @facan How to tell us about new URLs 1. Get your Bing Webmaster tools API ID https://www.bing.com/webmaster/api/ 2. Submit URLs online Example using wget wget.exe "https://ssl.bing.com/webmaster/api.svc/pox/SubmitUrl?apikey=7737def21c404dcdaf23bea715e61436" --header="Content-Type: application/xml; charset=utf-8" --post-data="<SubmitUrl xmlns="http://schemas.datacontract.org/2004/07/Microsoft.Bing.Webmaster.Api"><siteUrl>http://www.bing. com</siteUrl><url>http://www.bing.com/fun/?fabrice=pubcon</url></SubmitUrl>"
- 31. fabricecanel @facan Getting all content indexed lighting fast Yoast SEO enabling Bing Real Time indexing And soon elsewhere… • Your sites • Content Management System • Web hosting • SEO plug ins • SEO agencies
- 32. fabricecanel @facan Key takeaway … and the new magic Don’t forget the basic : TOP 3 SEO guidelines table Bing Webmaster Tools https://www.bing.com/webmaster/ AND Bing Webmaster APIs https://www.bing.com/webmaster/api/