SlideShare a Scribd company logo

Rank Dynamics Presentation

•Download as PPTX, PDF•
•
Mark Cramer
Mark Cramer

This document appears to be a presentation by Mark Cramer, CEO of Rank Dynamics, describing their technology for real-time contextualization of search results based on user behavior. It discusses tests Rank Dynamics conducted partnering with eBay that led to increased user engagement and sales. It also outlines the problems with current static search results and how Rank Dynamics' approach dynamically re-ranks results based on an inferred user intent model in real-time.

1 of 16
October 16th, 2015 Product & Company Presentation Mark Cramer CEO mark@rankdynamics.com
©2015 Rank Dynamics | All Rights Reserved 2 Vision Transform search into a dynamic experience where fluid result pages respond to user actions in real time
©2015 Rank Dynamics | All Rights Reserved 3 eBay Test #1: Splitting Traffic
©2015 Rank Dynamics | All Rights Reserved 4 eBay Test #1 (cont.): Substantial Increase in Sales Increased Engagement +32% increase in sales +30% increase in bids + sales +3.4% increase in clicks from SERP +8.2% increase in clicks from SERP beyond top 10 results Z-test for comparing two binomials: +33% improvement with 98.8% confidence (z-score -2.25)
©2015 Rank Dynamics | All Rights Reserved 5 eBay Test #2: Result Interleaving More relevant products +9% increase in CTR beyond top 10 results for queries with 200+ results +9% increase in CTR
©2015 Rank Dynamics | All Rights Reserved 6 Benefits Delivered Value Proposition Distribute technology to web as well as mobile searchers • Deliver technology to build “dynamic ranking” into shopping search • Integrate seamlessly into existing platforms, back-end systems and new product development Enhance retention and increase engagement Increase bids, sales and revenue • Boost RPS with real-time shopping contextualization
©2015 Rank Dynamics | All Rights Reserved 7 The Problem with Search A search can return thousands or millions of results. Current search engines: • Include irrelevant content in the results • Misinterpret users’ search terms • Have difficulty handling multiple or changing intents • Return predetermined, static result sets Users often have to dig through pages of results or reformulate their query in order to find what they need
©2015 Rank Dynamics | All Rights Reserved 8 Real-time Contextualization for Shopping Proprietary technology processes search results post-query to bring forward the content that is most relevant. Now. Fossil SEARCH Rank Dynamics surfaces the relevant search results based on real-time user actions.
©2015 Rank Dynamics | All Rights Reserved 9 Customized Subsequent Pages Navigating to a subsequent page will produce an instantly contextualized experience. Fossil SEARCH Works with traditional pagination as well as infinite scroll Boost Activity +33% increase in bids and sales
©2015 Rank Dynamics | All Rights Reserved 10 Instantaneous Relevancy & Re-ranking “On the Fly”Relevance Rank Results are targeted using instantaneous user intent model. Recalculated relevancies as determined using instantaneous user intent model InstantaneousRelevance ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 10
dolphins Subsequent result pages dynamically ranked in real-time! Real-time recommendations based on your activity ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 11 Organic Results Completely customized page two
Real-time Contextualization for Shopping The real-time inferred intent model is used to contextualize a shopping experience. Behavior signals will immediately produce a dynamic response, significantly facilitating the shopping experience. digital camera SEARCH 1 2 3 ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 12
Subsequent Shopping Pages Contextualized Navigating to the next Shopping page will produce an instantly contextualized page 2. Subsequent page contextualization can produce improvements in CTR beyond 40%. SEARCH 1 2 3digital camera ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 13
Real-time Contextualization for Image Search Real-time Contextualization with a grid layout. Behavior signals trigger immediate contextualization with grid results. Subsequent pages, even with infinite scroll, are contextualized in real-time. ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 14
©2015 Rank Dynamics | All Rights Reserved 15 The Team Mark Cramer CEO Mike Wertheim Chief Architect • Founder of Rank Dynamics • Over 20 years of technology industry experience, from engineer to executive • BS in Electrical Engineering from MIT • MBA from Harvard Business School • Over 20 years of software development experience • Content reviewer on the book "Bitter EJB," published by Manning Publications in 2003 • BS in Computer Engineering from Carnegie Mellon
©2015 Rank Dynamics | All Rights Reserved 16 Milestones 2008 2009 2010 2011 2012 2013 2014 Feb 2008 Launched browser extension Apr 2008 Closed $600K in seed funding Jan 2009 Favorably reviewed in the Mossberg Solution column of the WSJ Jul 2009 Research published by SIGIR: “Demonstration of Improved Search Result Relevancy Using Real-Time Implicit Feedback” Aug 2009 Launched search engine, achieved 1 million downloads and 100 million cumulative queries Jan 2012 Awarded patent “Dynamic Search Engine Results Employing User Behavior” Feb 2012 Awarded patent “Adaptive UI for Real-Time Relevance Feedback” Dec 2012 Surpassed 6 billion cumulative queries Sept 2011 Surpassed 10 million queries per day – 2.8 billion cumulative queries Apr 2013 Awarded patent “Real-Time Implicit User Modeling for Personalized Search” Dec 2014 5th patent issued

More Related Content

Rank Dynamics Presentation

  • 1. October 16th, 2015 Product & Company Presentation Mark Cramer CEO mark@rankdynamics.com
  • 2. ©2015 Rank Dynamics | All Rights Reserved 2 Vision Transform search into a dynamic experience where fluid result pages respond to user actions in real time
  • 3. ©2015 Rank Dynamics | All Rights Reserved 3 eBay Test #1: Splitting Traffic
  • 4. ©2015 Rank Dynamics | All Rights Reserved 4 eBay Test #1 (cont.): Substantial Increase in Sales Increased Engagement +32% increase in sales +30% increase in bids + sales +3.4% increase in clicks from SERP +8.2% increase in clicks from SERP beyond top 10 results Z-test for comparing two binomials: +33% improvement with 98.8% confidence (z-score -2.25)
  • 5. ©2015 Rank Dynamics | All Rights Reserved 5 eBay Test #2: Result Interleaving More relevant products +9% increase in CTR beyond top 10 results for queries with 200+ results +9% increase in CTR
  • 6. ©2015 Rank Dynamics | All Rights Reserved 6 Benefits Delivered Value Proposition Distribute technology to web as well as mobile searchers • Deliver technology to build “dynamic ranking” into shopping search • Integrate seamlessly into existing platforms, back-end systems and new product development Enhance retention and increase engagement Increase bids, sales and revenue • Boost RPS with real-time shopping contextualization
  • 7. ©2015 Rank Dynamics | All Rights Reserved 7 The Problem with Search A search can return thousands or millions of results. Current search engines: • Include irrelevant content in the results • Misinterpret users’ search terms • Have difficulty handling multiple or changing intents • Return predetermined, static result sets Users often have to dig through pages of results or reformulate their query in order to find what they need
  • 8. ©2015 Rank Dynamics | All Rights Reserved 8 Real-time Contextualization for Shopping Proprietary technology processes search results post-query to bring forward the content that is most relevant. Now. Fossil SEARCH Rank Dynamics surfaces the relevant search results based on real-time user actions.
  • 9. ©2015 Rank Dynamics | All Rights Reserved 9 Customized Subsequent Pages Navigating to a subsequent page will produce an instantly contextualized experience. Fossil SEARCH Works with traditional pagination as well as infinite scroll Boost Activity +33% increase in bids and sales
  • 10. ©2015 Rank Dynamics | All Rights Reserved 10 Instantaneous Relevancy & Re-ranking “On the Fly”Relevance Rank Results are targeted using instantaneous user intent model. Recalculated relevancies as determined using instantaneous user intent model InstantaneousRelevance ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 10
  • 11. dolphins Subsequent result pages dynamically ranked in real-time! Real-time recommendations based on your activity ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 11 Organic Results Completely customized page two
  • 12. Real-time Contextualization for Shopping The real-time inferred intent model is used to contextualize a shopping experience. Behavior signals will immediately produce a dynamic response, significantly facilitating the shopping experience. digital camera SEARCH 1 2 3 ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 12
  • 13. Subsequent Shopping Pages Contextualized Navigating to the next Shopping page will produce an instantly contextualized page 2. Subsequent page contextualization can produce improvements in CTR beyond 40%. SEARCH 1 2 3digital camera ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 13
  • 14. Real-time Contextualization for Image Search Real-time Contextualization with a grid layout. Behavior signals trigger immediate contextualization with grid results. Subsequent pages, even with infinite scroll, are contextualized in real-time. ©2015 Surf Canyon Incorporated dba Rank Dynamics | All Rights Reserved 14
  • 15. ©2015 Rank Dynamics | All Rights Reserved 15 The Team Mark Cramer CEO Mike Wertheim Chief Architect • Founder of Rank Dynamics • Over 20 years of technology industry experience, from engineer to executive • BS in Electrical Engineering from MIT • MBA from Harvard Business School • Over 20 years of software development experience • Content reviewer on the book "Bitter EJB," published by Manning Publications in 2003 • BS in Computer Engineering from Carnegie Mellon
  • 16. ©2015 Rank Dynamics | All Rights Reserved 16 Milestones 2008 2009 2010 2011 2012 2013 2014 Feb 2008 Launched browser extension Apr 2008 Closed $600K in seed funding Jan 2009 Favorably reviewed in the Mossberg Solution column of the WSJ Jul 2009 Research published by SIGIR: “Demonstration of Improved Search Result Relevancy Using Real-Time Implicit Feedback” Aug 2009 Launched search engine, achieved 1 million downloads and 100 million cumulative queries Jan 2012 Awarded patent “Dynamic Search Engine Results Employing User Behavior” Feb 2012 Awarded patent “Adaptive UI for Real-Time Relevance Feedback” Dec 2012 Surpassed 6 billion cumulative queries Sept 2011 Surpassed 10 million queries per day – 2.8 billion cumulative queries Apr 2013 Awarded patent “Real-Time Implicit User Modeling for Personalized Search” Dec 2014 5th patent issued