![A Knowledge Correlation Search Engine
How We Model Demand For N‐dimensional Search
We are interested first in current demand, and then in latent demand.
3
To model existing demand, we allocate the approximately 17% of searches [five, six, seven or more terms ] as
evidence of demand for N‐dimensional search. We are able to do this means of statistical analysis and phrase
composition analysis on such queries.
Demand for search should be characterized as a derived demand; i.e., households only demand search in the
course of carrying out other economic activities and not for the pleasure of search in and of itself. Household
demand for search is only determined by demand for exogenous economic activities. In some views, the cost
8
of search is considered virtually irrelevant, but search costs have an effect on consumer behavior . Search
time is the major component of variable costs experienced by those using major search engines to find
information on the internet.
Because of these characteristics, latent demand for search can therefore be modeled in the much same
5
manner as latent demand for transportation, called “induced travel” .
The underlying theory behind induced travel is based upon the simple economic theory of supply and demand.
5
Any increase in highway capacity (supply) results in a reduction in the time cost of travel . Like travel, search
5,8
(in a characteristic shared by all internet applications) is highly sensitive to time cost and those costs have
been calculated. In search, we construe increased capacity to mean reduction in time to reach a relevant
5
search result. When any good (travel or search) is reduced in cost, demand for that good increases . Induced
6
travel that represents new trips and longer trips is called “generative” . On average, every 10% reduction in
7
travel time (cost) produces an 8% increase in travel . Reduction in search costs have been found to produce
8,9
increases in effective demand elasticity for consumer goods . Although we estimate all increase in “capacity”
to be attributable to N‐dimensional Search, to be conservative, we estimate a price elasticity of only ‐.5.
Further, 80% to 100% of all new capacity in a transportation system is absorbed by induced travel over time,
and the relative near and long term effect of improved search has been shown to exhibit similar
8
characteristics .
27](https://image.slidesharecdn.com/knowledgecorrelationsearchengine-131101135149-phpapp01/85/The-N-Dimensional-Search-Engine-27-320.jpg)
![A Knowledge Correlation Search Engine
Citations
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Crane, J. 2007. August Search Market Share. Compete.com
http://blog.compete.com/2007/09/12/search‐market‐share‐august/
Nielsen//NetRatings MegaView Search, October 2006 (as reported by Metrics2.0, November 2006). Sizing Web Search
Search Marketing: Key Stats, Forecasts & Metrics.
http://www.metrics2.com/blog/2006/11/15/sizing_web_search_search_marketing_key_stats_forec.html
OneStat.com (as reported by Metrics2.0, November 2006). Two word search phrase is most popular
http://www.metrics2.com/blog/2006/10/17/two_word_search_phrase_is_most_popular_one_word_se.html
Meeker, M., Joseph, D.A, 2006. US Internet Advertising Outlook, 2006‐2010E. Morgan Stanley Research N.A.
Noland, R.B., 1999. Relationships between highway capacity and induced vehicle travel. (Published) Transportation
Research Part A 35 (2001) 47‐72
Cervero, R., 2001. Road Expansion, Urban Growth, and Induced Travel: A Path Analysis. University of California, Berkeley
Litman, T., 2007. Generated Traffic and Induced Travel. Victoria Transport Policy Institute
Ghose, A., Gu, B., 2006. Search Costs, Demand Structure and Long Tail in Electronic Markets. Stern School of Business
Levin, A.T., Yun, T., 2007. Customer Search, Monopolistic Competition, and Relative Price Dispersion. Board of Governors of
the Federal Reserve System
Gao, M., Hyytinen, A., Toivanen, O. 2006 Demand for and Pricing of Mobile Internet
Dave Girouard, 2006. Enterprise Search October, 2006 (Lecture) Austin, TX
Buteney, R.J., 2007 Description of Greater Than 4 Term Search. Telephone Interview. Compete.com
Haroon, A. 2006. How to Create An Effective Paid Search Campaign. Direct Marketing News [citing Hitwise]
http://www.dmnews.com/cms/dm‐news/search‐marketing/37779.html
Doyle, J. 2006. How to Capture Your target Consumers' Profile Data Online. Pharmaceutical Executive
http://www.pharmexec.com/pharmexec/article/articleDetail.jsp?id=301760
Lipsman, A. 2007. 61 Billion Searches Conducted Worldwide in August.. comScore
http://www.comscore.com/press/release.asp?press=1802
Gartner, J. 2006. Scrutiny on the Bounty. Revenue http://www.revenuetoday.com/story/scrutiny‐on‐the‐bounty
Contact: Mark Bobick m.bobick@correlationconcepts.com 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved.
Proprietary and Confidential. Dissemination without permission strictly prohibited
31](https://image.slidesharecdn.com/knowledgecorrelationsearchengine-131101135149-phpapp01/85/The-N-Dimensional-Search-Engine-31-320.jpg)
More Related Content
Viewers also liked (11)
Similar to The N-Dimensional Search Engine (20)
More from s0P5a41b (10)
Recently uploaded (20)
The N-Dimensional Search Engine
- 1. CAPTURING THE UNSERVED WEB SEARCHERS’ MARKET THE N-DIMENSIONAL SEARCH ENGINE Make Sence, Inc.
- 2. OUTLINE • State of the Web: Current Search Engines produce unsatisfactory results for complex searches • The Opportunity: Search Engine limited capabilities = unserved users & unrealized revenues • The Solution: Solve complex queries using the N-Dimensional Search Engine • The Benefits: Access the unserved 16% web search market and corresponding advertising opportunities • The Technology: Knowledge Correlation • The Make Sence Group: Our team and product history 2
- 3. THE STATE OF WEB SEARCHES Current search capability is limited IF: Simple question ( “green tea”) + asked enough times by enough people + specific documents on the “open” WWW THEN: search works = “Limited Search” Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 3
- 4. THE STATE OF WEB SEARCHES Limited Searches work for approximately 84 ‐ 85% of the 265 million daily searches on major search engines1,2,3.11 Using Limited Searches, the industry collected context sensitive 4 advertising revenue of US$177 per US household in 2005 Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 4
- 5. THE STATE OF WEB SEARCHES However, Limited Searches provide unsatisfactory results for other queries (1) (2) (3) (4) (5) Complex questions Questions never asked before Questions previously asked infrequently Questions with answers spread piecemeal thru many documents Questions whose answers are not part of the open Web (the “Complex Queries”) Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 5
- 6. THE STATE OF WEB SEARCHES The industry response of “Better Limited Searches” does Not satisfy Complex Queries requirements. _______________ •Latent Semantic Analysis •Topic Detection •Concept Extraction •Document Retrieval •User Profiling •Interactive Dialog with User •GUI Enhancements •Search Integrated Portals Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 6
- 7. THE OPPORTUNITY Google has captured the general Limited Search market both technologically (“page rank” algorithm etc.) & using its business models (AdWords and AdSense) Vertical search and niche providers , & big footprint web players, are taking conversion rate traffic‐ using Limited Searches BUT – the Complex Queries market remains unserviced!! Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 7
- 8. THE OPPORTUNITY A new market exists for Search advertisers who currently: 1. Are paying for ads which are inappropriately placed in front of Complex Queries searchers “ out of context “ due to unsatisfactory search results OR 2. Do not get their ads placed in front of Complex Queries searchers who would be potential consumers ! Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 8
- 9. THE OPPORTUNITY Scale of the Opportunity: Approximately 16% of current search queries are likely to be Complex Queries!! Exploiting this Opportunity: Such Complex Queries demand a new and “more complete search” – satisfied only by fundamentally extending Search Engine capabilities Losing this Opportunity: Simply “Improving” Limited Searches will NOT address this major opportunity An “N‐dimensional Search” is required to exploit this opportunity Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 9
- 10. THE SOLUTION: N‐DIMENSIONAL SEARCHES Make Sence, Inc. has developed The Knowledge Correlation Search Engine (CSE) to capture N‐Dimensional Searches…………. .........and to help a search provider capture the 16 to 24% high value Complex Query search user market!! Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 10
- 11. THE SOLUTION: N‐DIMENSIONAL SEARCHES What value is an N‐dimensional Search? Used with a Limited Search Engine, an N‐Dimensional Search provides a more complete search capability to meet the needs of all of today's search engine users Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 11
- 12. THE SOLUTION: N‐DIMENSIONAL SEARCHES When is an N‐dimensional Search Used? Appropriate for virtually all Complex Queries, including when search terms are: •Semantically complex and/or compound •Have a minimum of two terms, phrases or concepts •Contain “content words” which exhibit lexical and/or semantic overlap or disjunction •Usually are comprised of 5, 6, or 7 terms, phrases or concepts Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 12
- 13. THE SOLUTION: N‐DIMENSIONAL SEARCHES Comparing a Limited Search with the Knowledge Correlation Search Engine (CSE) Decision Index Document Knowledge Decomposition Rank New Knowledge Generation & Action SE Make Sence X Y SE SE CSE X Y CSE X CSE Y X Y Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 13
- 14. THE SOLUTION: N‐DIMENSIONAL SEARCHES Comparing a Limited Search with the Knowledge Correlation Search Engine (CSE) Single Term Simple Search Query SE Multi-Term Simple Search Query N-dimensional Search Query SE Make Sence CSE X Y CSE X Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited Y 14
- 15. THE SOLUTION: N‐DIMENSIONAL SEARCHES Sample Knowledge Correlation Search Result Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 15
- 16. THE BENEFITS OF N‐DIMENSIONAL SEARCHES Why Should Providers Offer N‐dimensional Search capability? Multiple Markets 1. New Market: N‐Dimensional search capability opens up a new unserved/underserved space in the Search Market; 15–16% of current searches (N‐dimensional) can now be satisfied 2. Latent Market: consists of 7–8% of existing search query volume is available for capture (latent demand) 3. Retention of Existing Market: Opportunity to generate higher overall site loyalty among focused users Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 16
- 17. THE BENEFITS OF N‐DIMENSIONAL SEARCHES Why Should Providers Offer N‐dimensional Search capability? The Monetization opportunity from high value demographics! Profile : N‐Dimensional search query users are among the most focused4,13 , more likely to click thru and “stick” and to “take the next step” of submitting contact information or making purchases Market size : 32 million N‐Dimensional search query users a day get no answers and receive no useful ads! 15 million additional search customers15 are now available Market Value: N‐Dimensional search query users create a market for premium ad fees Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 17
- 18. THE BENEFITS OF N‐DIMENSIONAL SEARCHES Why Should Providers Offer N‐dimensional Search capability? To solve the major problem of acquiring User Profiles. •The more specific User Profile information about individual consumers the more valuable a sales lead becomes16 •N‐dimensional Search takes even an anonymous user’s input and determines the user’s actual intent for the search •With N‐dimensional Search, advertisers can get the same results as with expensive customer capture methods16 Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 18
- 20. A Knowledge Correlation Search Engine Knowledge Correlation is a novel mechanism for capturing the associations hidden in data There are two forms of Knowledge Correlation: Connect the Dots X Free Association Y ? ! Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 20
- 21. A Knowledge Correlation Search Engine Connect the Dots X Y Used to find if ANY association can be constructed connecting a given origin to a given destination X TERM = “AUTOMOBILES” Y TERM = “POLLUTION” set of terms, phrases or concepts: AUTOMOBILES CONTAIN ENGINES ENGINES BURN FUEL FUEL BECOMES EMISSIONS EMISSIONS ARE POLLUTION Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 21
- 22. A Knowledge Correlation Search Engine Free Association ? ! Has this happened to you? •Man is seated at a table in a restaurant •He notices another couple, strangers, sitting nearby •He further notices that the woman is wearing a red dress •He thinks to himself: •“My wife has a red dress like that” •“She last wore that dress when we went to the Bahamas three years ago” •“My buddy George told me a couple of months back he was going to the Bahamas. I wonder if he went.” •“My wife is friends with George’s wife (Carol)” •Man asks his wife, “Have you spoken to Carol lately?” In a fraction of a second, the man in this example followed a chain of memory (knowledge) fragments and acted upon them. Humans think like this all the time. Now, using Knowledge Correlation Technology, computers can emulate this type of human experience‐driven association. Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 22
- 23. A Knowledge Correlation Search Engine Applications for Connect the Dots: •Risk Mitigation •Compliance (Sarbanes Oxley) •Workforce Re‐engineering •Anti‐Money Laundering Applications for Free Association: X Y ? ! •Compound Discovery (Bio‐Pharma) •Legal Precedent Discovery •News and Event Analysis •Uncertainty Modeling Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 23
- 24. A Knowledge Correlation Search Engine How does Knowledge Correlation work? 1. Take a lot of documents 2. Decompose the documents into “fragments of knowledge” and store all the fragments in a large database 3. Try millions and millions and millions of ways to chain the fragments together to answer a user question. AUTOMOBILES CONTAIN ENGINES ENGINES BURN FUEL FUEL BECOMES EMISSIONS ARE EMISSIONS POLLUTION XX XX X YY YY Y If even one chain from X to Y can be constructed… Knowledge has been Discovered! The chains are “correlations” 24
- 25. A Knowledge Correlation Search Engine How is Knowledge Correlation used in Search? Based upon the user’s input, all the chains (correlations) that can be constructed linking X to Y create a network, like the one below… X Y Each “knowledge fragment” in the network still “points back” to the document from which the “knowledge fragment“ was extracted… By linking user input to fragment to document, user input is linked to documents… 25
- 27. A Knowledge Correlation Search Engine How We Model Demand For N‐dimensional Search We are interested first in current demand, and then in latent demand. 3 To model existing demand, we allocate the approximately 17% of searches [five, six, seven or more terms ] as evidence of demand for N‐dimensional search. We are able to do this means of statistical analysis and phrase composition analysis on such queries. Demand for search should be characterized as a derived demand; i.e., households only demand search in the course of carrying out other economic activities and not for the pleasure of search in and of itself. Household demand for search is only determined by demand for exogenous economic activities. In some views, the cost 8 of search is considered virtually irrelevant, but search costs have an effect on consumer behavior . Search time is the major component of variable costs experienced by those using major search engines to find information on the internet. Because of these characteristics, latent demand for search can therefore be modeled in the much same 5 manner as latent demand for transportation, called “induced travel” . The underlying theory behind induced travel is based upon the simple economic theory of supply and demand. 5 Any increase in highway capacity (supply) results in a reduction in the time cost of travel . Like travel, search 5,8 (in a characteristic shared by all internet applications) is highly sensitive to time cost and those costs have been calculated. In search, we construe increased capacity to mean reduction in time to reach a relevant 5 search result. When any good (travel or search) is reduced in cost, demand for that good increases . Induced 6 travel that represents new trips and longer trips is called “generative” . On average, every 10% reduction in 7 travel time (cost) produces an 8% increase in travel . Reduction in search costs have been found to produce 8,9 increases in effective demand elasticity for consumer goods . Although we estimate all increase in “capacity” to be attributable to N‐dimensional Search, to be conservative, we estimate a price elasticity of only ‐.5. Further, 80% to 100% of all new capacity in a transportation system is absorbed by induced travel over time, and the relative near and long term effect of improved search has been shown to exhibit similar 8 characteristics . 27
- 28. A Knowledge Correlation Search Engine Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 28
- 29. A Knowledge Correlation Search Engine Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 29
- 30. A Knowledge Correlation Search Engine Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 30
- 31. A Knowledge Correlation Search Engine Citations 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. Crane, J. 2007. August Search Market Share. Compete.com http://blog.compete.com/2007/09/12/search‐market‐share‐august/ Nielsen//NetRatings MegaView Search, October 2006 (as reported by Metrics2.0, November 2006). Sizing Web Search Search Marketing: Key Stats, Forecasts & Metrics. http://www.metrics2.com/blog/2006/11/15/sizing_web_search_search_marketing_key_stats_forec.html OneStat.com (as reported by Metrics2.0, November 2006). Two word search phrase is most popular http://www.metrics2.com/blog/2006/10/17/two_word_search_phrase_is_most_popular_one_word_se.html Meeker, M., Joseph, D.A, 2006. US Internet Advertising Outlook, 2006‐2010E. Morgan Stanley Research N.A. Noland, R.B., 1999. Relationships between highway capacity and induced vehicle travel. (Published) Transportation Research Part A 35 (2001) 47‐72 Cervero, R., 2001. Road Expansion, Urban Growth, and Induced Travel: A Path Analysis. University of California, Berkeley Litman, T., 2007. Generated Traffic and Induced Travel. Victoria Transport Policy Institute Ghose, A., Gu, B., 2006. Search Costs, Demand Structure and Long Tail in Electronic Markets. Stern School of Business Levin, A.T., Yun, T., 2007. Customer Search, Monopolistic Competition, and Relative Price Dispersion. Board of Governors of the Federal Reserve System Gao, M., Hyytinen, A., Toivanen, O. 2006 Demand for and Pricing of Mobile Internet Dave Girouard, 2006. Enterprise Search October, 2006 (Lecture) Austin, TX Buteney, R.J., 2007 Description of Greater Than 4 Term Search. Telephone Interview. Compete.com Haroon, A. 2006. How to Create An Effective Paid Search Campaign. Direct Marketing News [citing Hitwise] http://www.dmnews.com/cms/dm‐news/search‐marketing/37779.html Doyle, J. 2006. How to Capture Your target Consumers' Profile Data Online. Pharmaceutical Executive http://www.pharmexec.com/pharmexec/article/articleDetail.jsp?id=301760 Lipsman, A. 2007. 61 Billion Searches Conducted Worldwide in August.. comScore http://www.comscore.com/press/release.asp?press=1802 Gartner, J. 2006. Scrutiny on the Bounty. Revenue http://www.revenuetoday.com/story/scrutiny‐on‐the‐bounty Contact: Mark Bobick [email protected] 702.882.5664 Copyright 2007 Make Sence Florida, Inc. All Rights Reserved. Proprietary and Confidential. Dissemination without permission strictly prohibited 31