Search Meets Blogs and Wikis Feb 8 2007

Search Meets Blogs and Wikis Feb 8, 2007 San Diego Greg Lloyd, President and co-founder Traction Software Inc Providence, Rhode Island www. Traction. Software. com

Managers Say the Majority of Information Obtained for Their Work Is Useless, Accenture Survey Finds NEW YORK and LONDON – Jan. 4, 2007 – Middle managers spend more than a quarter of their time searching for information necessary to their jobs, and when they do find it, it is often wrong, according to results of an Accenture (NYSE: ACN) survey released today… Managers spend up to two hours a day searching for information, and more than 50 percent of the information they obtain has no value to them… Nearly three out of five respondents (59 percent) said that as a consequence of poor information distribution, they miss information that might be valuable to their jobs almost every day because it exists somewhere else in the company and they just can not find it… In addition, 42 percent of respondents said they accidentally use the wrong information at least once a week, and 53 percent said that less than half of the information they receive is valuable… 40 percent of respondents said that other parts of the company are not willing to share information, and 36 percent said there is so much information available that it takes a long time to actually find the right piece of data.

Euan Semple: Survey proves 90% of managers are clueless … this survey that rang many bells for me and brought back my brief period in charge of the intranet at the BBC. I can so well remember the frustration when managers would say “I can never find anything on the intranet” as if it was my fault. On my grumpier days I would respond with “Have you ever tried sharing anything on it yourself” but mostly I bored them into submission with explanations about why intranet search couldn’t work like Google as this was basically what they thought they wanted. … Like Jack I came to believe that what people really wanted was to find someone who knew what they were talking about. Even if that “knew what they were talking about” meant knowing which document to read, why and where it was to be found. So what we did was start building online social spaces like forums, blogs and wikis in which highly contextual, subjective, complex patterns and information could start to surface about anything and everything in the business that was interesting and worth writing about. The result was that when someone said on our forums “I need to find the official documentation on x because I am about to do y” they were usually rewarded, and very quickly, with multiple answers along the lines of “Well I found this document answered my questions because …. . ” pointing them at the documentation. Indeed increasingly the source they were directed to was a blog or a wiki containing up to date, contextualized information. Euan Semple FASTforward 07 blog 8 Jan 2007

Foraging Theory If you were a rational bird in winter … How much time would you spend gathering for food where you last found it? You know how long it takes to get there You know how much you can gather in an hour You know how crowded it is You know it will eventually become harder to find food there Versus flying to discover a new patch? Which might be richer Which might be less crowded But costs energy to find And costs more energy if you need to travel farther Birds who do this well survive and prosper. Birds who don’t - die!

Managers are not rational birds, but … Information Foraging theory adapts optimal foraging models to study people finding information rather than animals finding food Where do people look for information ? How effective are they in finding what they need ? How can we create (human as well as computer) systems that help? “The basic hypothesis of Information Foraging Theory is that, when feasible, natural information systems evolve towards stable states that maximize gains of valuable information per unit cost …” - Pirolli and Card, Jan 1999

One Strategy - Find the right “information patch” An “information patch” is a collection of information that you or others have created which is likely to contain relevant information Examples: The pile of papers on your desk - messy but context filtered Your email slush pile - messy but context filtered A formal repository - best formal work products A person - ideally an expert or stakeholder

Blogs and Wikis are Enriched Information Patches “One may also enrich information patches by using filtering processes. For instance, people often filter their readings on a topic by first generating and filtering bibliographic citations and abstracts. Many computer systems for electronic mail, news, and discussion lists now include filters. Such enrichment activities create the trade-off problem: should one continue to enrich patches to improve future within-patch foraging or should one turn to exploiting them? ” - Pirolli and Card Jan 1999 “One can also invest time creating a new patch or enriching an old one. A Journal (or Memex) is a tool that makes it economical to create, cultivate and reduce the cost of finding your patches. Unlike food, old information is not 'consumed' - it can become more valuable as it is cultivated. ” - Greg Lloyd’s Journal Aug 1999 Traction 7825

Blogs and Wikis meet Enterprise Search Deploy shared spaces centered on a business purpose Wiki - Build out a space page by page Group blog - Log notes, meetings, questions, proposals, decisions Personal blogs - general notes and comments Attach working files to the shared space File attachments to blog / wiki pages Files in web folders available per shared space (project) This has several beneficial effects: Working communication is better Blog / wiki content creates an “authored index” The blog space is a natural place to ask questions in context And - FAST search can use blog/wiki content to contextualize search

A Wiki is like a House

Built by many people

A Blog is like the record of building the house …

Business needs the blog and wiki views working together The product you wish to build, sell and support The external intelligence, dialog, and decisions of teams working together over time to create and sustain the product

Business Purpose Competitive Intelligence Sales Communication Research and Development Journals Program / Project Communication Corporate Communication and Financial Reporting Exception tracking Manufacturing and operations log Blogs and wikis create shareable, searchable and secure spaces Security makes it possible to open selected spaces to specific external stakeholders (customers, suppliers, resellers, PR and other advisors), while giving internal stakeholders a global view

Douglas Engelbart - Hypertext for High Performance Teams 1968 NLS (o. N Line System). A Hypertext Journal for high performance teams 1992 Towards High Performance Organizations: A Strategic Role for Groupware

Dialog, External Intelligence, Knowledge Product

Journal + Web = Searchable Context • • Stable content Paragraph addressable Built-in time order Flexible viewspecs for assembly of content • • Web browser interface Web linkable Easily authored Each server creates a cluster of secure spaces

Website & Intranet: Integrated Justice Information Systems Institute Traction powered IJIS website: Now easy to edit, also Includes Search, RSS, and member Login The Payoff: About 400 people working on 37 projects now collaborate virtually. One committee cut by half the number of phone and in-person meetings it holds. Traction driven: Website Community Extranet Staff Intranet Provides out of the box: Blog & Wiki per Group Integrated Search Dynamic RSS/ATOM Threaded Discussion Document Management Access Control

Competitive Intelligence: Ipsen Pharmaceutical Readership ramped quickly to 8, 000 to 14, 000 page views per month Results • • Federated collection of 6 blogs hold over 3, 700 articles 640 comments A community and repository for knowledge of the business environment Core team of 250 named accounts (Intranet, Extranet, email), Available to 3, 500 employees (Intranet) Read and comment Dialogue around information is increasing (within the blog and in meetings) Contributed to making CI important to business decision making Market-relevant communication is moving from email to the blogs

Integration Example – QL 2 Software Automatically capture web site changes and search results Publish and interact with the content in Traction. CI Staff can use inline comments and labels to analyze and decide what to post for broad use. Respond in hours or minutes - not days or weeks.

Traction Team. Page FAST Search Module FAST Search and Content Navigation Automatic Entity and Keyword extraction and drill down provides context sensitive content navigation - over and above integrated search. Search results and drill down navigation are permission filtered on the fly.

Search controls

Keywords from search results

Narrow results by keyword

Narrowed results

Label Navigation

Project Navigation

Comments - Most Recent First

Narrow by name - based on automatic name extraction

References Accenture Newsroom: Managers Say the Majority of Information Obtained for Their Work Is Useless, Accenture Survey Finds - Accenture, 4 Jan 2007 Survey proves 90% of managers are clueless - Euan Semple FASTForward 07 blog, 8 Jan 2007 Information Foraging Theory: Adaptive Interaction with Information (Oxford Series in Human-Technology Interaction) - Peter Pirolli, Feb 2007 Information Foraging - Peter Pirolli, Stuart Card, Xerox PARC UI Research Group, Published in Psychological Review Jan 1999

Traction Software Inc Greg Lloyd President and Co-Founder grl@tractionsoftware. com (401)-528 -1145 Traction Software, Inc. 245 Waterman Street Suite 309 Providence, RI 02906 USA www. Traction. Software. com

At a Glance: Traction Software Inc Market Focus • Secure, scalable Blog and Wiki software for enterprise use Funding/Advisors • Two founders with 5 decades of hypertext experience • Funding: In-Q-Tel (venture arm of the CIA), Slater Interactive and private investors • Advisor: Professor Andy van Dam – Brown University, Microsoft Research Advisory Board, Hypertext pioneer Business Highlights • • • Over 220 enterprise customers by 2006 Global 1000 and Government focus in US, Japan, Europe Strong positions in pharma, finance, manufacturing, government Building strong partnerships: FAST Search & Transfer, QL 2 Software “Enterprise 2. 0” is sparking customer and channel interest Award Winning Product and Market Leading Position • • Best Enterprise Current Release: 3. 7. 2 Info. World 2007 Technology of the Year - Best Enterprise Wiki Industry Recognition: Top 100 Awards from KMWorld, e. Content, Red Herring Top Reviews: Burton Group, Forrester Research, Info. World, Release 1. 0, Intranet Journal, Federal Computer Week, Fuld & Company and more. Wiki

Traction Team. Page Customer Summary Over 220 Customers • Over fifteen Fortune 500 Customers • Over ten Global 100 Customers in US, Japan, Europe • Five of Top Ten Global Pharma Customers Addressing over 1, 000 Enterprise Eyes Team. Page combines the group editing of a wiki and time-order of a blog with a unique comment and permission model Initial applications: Competitive Intelligence, Issue tracking, Product Management Pharma, Banking, Government are early adopters Emerging applications: Product, Process, Customer and Supply chain communication Pharma, Software, Electronics, and Manufacturers of engineered products Consultant / client communication Corporate communication / SEC financial reporting

Traction Team. Page | Info. World 2007 Technology of the Year Much sets Traction Team. Page apart, starting with the company’s core hypertext engine. This technology handles all the indexing, cross-reference, and navigational elements of Team. Page, so you can build extremely large enterprise wikis, yet quickly organize all the content. Traction has excellent usability together with the best — and most flexible — look and feel of the products tested. … Traction Team. Page clearly placed above the others with its superior ease of use and query ability that presented each user with just the knowledge they need. Wikis evolve as collaboration tools Info. World Test Center Review, Mike Heck, January 5, 2007 Best Enterprise Wiki

Beyond blogs and wikis, there’s Traction Although Traction's products are marketed as “enterprise weblog” offerings, they are in fact powerful hypertext systems that exploit blog- and wiki-related advances to maximize simplicity but also do so on a scale that addresses elaborate hypertext application scenarios that would overwhelm wiki alternatives. Blogs, Wikis and Beyond: New Alternatives for Collaboration and Communication Peter O’Kelly, Senior Analyst, Burton Group, March 2005 [Traction Team. Page] comes closest to bringing the visions of hypertext pioneer Doug Engelbart to fruition, and that it is also a very useful leading indicator in terms of features other vendors will eventually add. Hypertext and Compound/Interactive Document Models: Collaboration and Content Management Implications Peter O’Kelly, Senior Analyst, Burton Group, October 2006

The Traction Platform – What’s Different? Collaborative hypertext journal for Individuals and Groups working over time • “Hypertext Journaling” system maintains best of breed audit trail • Robust mechanism for handling edits and cross references over time • Journal made up of “project” spaces • Each project is analogous to a blog or a wiki space • Permissions are defined for the projects • Using standard Access Control Lists • Integrated with Active Directory and LDAP • Paragraph level cross project commenting, linking and labeling (social tagging) • An article in one project can be labeled “FAQ” or “Important” in another project - Labels also expose content from one project to another • Direct a comment on an article in Project A to a more private Project B • Make bi-directional links between articles or comments regardless of project • Comments, Labels and Links are permission filtered • Traction Views • Deliver any content slice or set of slices and • Are filtered based on sum of your permissions as well as any query • Can be subscribed to via RSS • Skins control look & feel, and function • Traction delivers different skins to different people (or software agents) at the same time

Traction Team. Page – Architecture Diagram

Traction Team. Page – Technical Overview • Java/J 2 SE Application • Simple installers for Windows, Solaris, Linux, MAC OSX • No dependencies (e. g. CGI, Perl, PHP, My. SQL, etc. ) • Journal records all changes (articles, comments, label changes, emails sent out, etc. ) • Articles have automatic edit history with Wikimedia style differences view • Web. DAV manages attachments to record file version history • Journal indexed for optimized for view delivery • Security and Authentication • Built-in user management, or connect to Active Directory, LDAP, or others via plug-in • Supports TLS and TLS with X. 509 certificates, NTLM, and LDAP Authentication • Access Control List driven permissions • Fine grained permission controls (Author, Comment, Read Own, Edit Own, Erase Own, Change Labels, Make Labels, Project Admin, etc. ) • Built-in Web. DAV web folders included in permissioned access model • Internationalized interface + Unicode (UTF-8) • Localized for Japanese, German, French; Supports local vocabulary overrides (e. g. Surfer dude) • SDK and SDL (skin definition language) allow for customization • Create new skins, views, and reports — HTML, XML, CSV, text, etc. • User Preferences • “Skins” are basis for most views, are selectable and customizable by user preference • Time zone preferences

Traction Team. Page – Inputs • Traction Instant Publisher • • Optional. NET client for Windows desktop and Internet Explorer integration Internet Explorer “blog selection”, or drag and drop capture of text, images, and back link to source Drag/drop attachments Built-in screen capture tool • XML-RPC and SOAP • Use most blog publishing tools to post to Traction via XML-RPC • Perl Traction publish script posts via SOAP • Zero Footprint Rich Text Editor • For new posts, edits, comments • Works in Mozilla, Firefox, Internet Explorer, Safari • Widgets can be produced quickly and published with a widget tool • Custom forms • Create custom forms with SDL and Java. Script • Can take advantage of the rich text editor • Email • Projects can read a POP 3 or IMAP mailbox; SSL and STARTTLS supported for all connections. • Author by email permission per project • Rules-based automatic label application (labels from any project) • With Traction Team. Page Feed Reader module: • Read and publish RSS / Atom feeds into Traction • Rules based automatic label application (labels from any project)

Traction Team. Page – Outputs • Skins and Views • • Standard views include: front page, news page, topic navigation with extensible view types Simple skin for PDAs and Phones Search skin for spiders RSS/ATOM skins for aggregators and data mining tools • Notification • RSS/ATOM – Best of breed implementation - Secure, Authenticated, Dynamic (based on permission filter and search criteria) - Subscribe to anything you can imagine searching for • Automated Email Newsletter - Permission filtered, sent at regular interval, including a paragraph from new articles and comments - Uses a newsletter (digest) skin which can be customized • Notifiers - Email and Jabber notifiers – notification of all new content, edits, label changes - Extensible - add your own • Collect Content and Send • Email – Users can select any set of content and mail out - Audit trail remembers who sent what to who, when • PDF / Word. ML – Users can select any set of content and generate a PDF or Word. ML Briefing Book - Can automatically create an Appendix with N levels of referenced articles • Traction Java SDK and SDL (Skin Definition Language) • Java SDK for extending SDL, Authentication and Directory, Notification, and Action