Projects

Iknow conducted a comprehensive review of the “top 10” commercial products. Information was collected from vendors’ product literature, independent product evaluations by market research firms such as Gartner, interviews with State of Michigan staff about their experiences with the products, and Iknow’s proprietary product database.

As a first step, Iknow compiled a list of EDM platform features and attributes to serve as the basis for the cross-platform comparisons. This list was created based on interviews with Michigan staff to understand the desired product functionality for their specific use cases. Iknow then performed a comprehensive analysis of features and functionality for ten market-leading products and the three systems most widely used within the State. Iknow incorporated data from the manufacturers’ product literature, expert reviews, and interviews. We also used literature and expert sources to analyze long-term file format storage options and made recommendations on alternatives.

The key deliverables from this project were:

• Analysis of features and functionality for the latest versions of the three primary EDM products currently in use within the State of Michigan’s departments and agencies.
• Analysis of features and functionality for the “top 10” commercial EDM products.
• Description of the functionality gaps between the existing products and the “top 10” commercial products and an evaluation of their importance.
• Analysis and description of alternatives file formats for long-term document retention.

Iknow began this assignment by reviewing a representative data subset and developing the categories for classifying the comments. Working with the company’s marketing, digital commerce, and business intelligence staff, more than 130 categories were ultimately selected. Next, the Analyze learn-by-example (LBE) capability was used to automate the processing of text comments. The software is able to read and analyze each sentence, determine its meaning and sentiment, and classify it into one of the defined categories.

Iknow worked with the company to test and validate the results from the automated classification and to establish the internal reporting process. Results were routed to specific executives throughout the company so that they could address the problems that fell within their spans of control.

The project consisted of the following six major work activities:

1. Content Source Analysis. Iknow worked with company’s Supply Chain Subject Matter Experts (SMEs) to identify the specific types of information that should be collected and displayed, and the specific sources of that information. Iknow synthesized the information and created a data element-data source map. Iknow evaluated alternate sources and prepared recommendations on the specific external sources of information that should be acquired for the SI Platform.

Iknow developed standard data element-naming conventions and definitions to assist in aggregating data from multiple sources. Iknow also created business logic for resolving conflicting data.

2. Supplier Master Index (Key) and the Data Workflow and Processing Algorithms. Iknow developed a master key/indexing strategy for uniquely identifying each potential supplier. We created a relational database that integrated the company’s internal supplier naming conventions, D&B® D-U-N-S® Numbers, official company names, North American Industry Classification System (NAICS) industry codes, and the names of product categories and commodity types. Iknow then designed the end-to-end process workflow for data acquisition, cleansing, matching, and disambiguation.

Note: This work stream was very important because each new piece of incoming data needed to be matched correctly to a unique company. While D-U-N-S Numbers are used to uniquely identify more than 100 million businesses worldwide, we estimated that up to 30 percent of the potential suppliers in the company’s SI database that are located outside the United States did not have a D-U-N-S Number. Additionally, NAICS codes are likely only available for companies operating in North America. Therefore, creating and updating a supplier master index and the supporting logic algorithms to resolve matches of companies with similar names was fundamental to the overall integrity of the system.

3. Intelligent Search (IS) Broker Creation. Iknow created the custom software code for acquiring the supplier data from the external sources. Iknow used the SAP BusinessObjects Intelligent Search product and built custom software brokers to access the specific websites and databases.
    
4. Text Analysis Rule Creation. Iknow developed a set of custom rules for the SAP BusinessObjects Text Analysis product. The custom rules are used to finely tune the natural language processing (NLP) algorithms so that only the desired entities, or data elements, are extracted and stored during the processing of incoming, unstructured text content.
    
5. User Experience Design and Development. Iknow developed high quality, user-friendly interfaces for the Supplier Intelligence application. The interfaces were web-based, multipage displays that contained advanced functionality for searching and displaying structured and unstructured data. Iknow designed the search, advanced search, faceted navigation of search results, and other search-related functionality. Iknow developed several display dashboards that present supply chain-related data in pie charts, bar charts, time-series plots, maps, and other display formats.
    
6. Business Logic and Technical Architectures. Iknow designed, developed, and tested the new Supplier Intelligence Platform following the company’s DoD-certified software development methodology. The Platform was implemented using a services oriented architecture (SOA). Core technical functionality and business logic included scheduling the raw data acquisition, performing data quality and data reconciliation activities, and loading the data into structured databases. Iknow also designed and developed an overall user experience that fosters interaction with an exploration of the data and provides the mechanisms for collecting human intelligence and other forms of feedback.

The Supplier Intelligence platform was implemented using the following software tools from SAP:

• SAP BusinessObjects Intelligent Search Language Processing
• SAP Business Objects Text Analysis (language: English)
• Text Analysis Language Processing (all other languages)
• SAP BusinessObjects DQM
• SAP BusinessObjects BI Package
• SAP Business Suite Business Expert Upgrade User
• SAP NetWeaver Enterprise Search
• SAP Portal

Other commercial, off-the-shelf (COTS) software products included Oracle databases and SAP ERP systems. The production system is hosted in the company’s secure data centers. The Platform and the supporting databases are accessible behind the company’s firewall.

*          *          *

Some of the project’s major deliverables included:

• Software installation in development, testing, and production environments
• Business requirements document
• Functional requirements document
• UI design and implementation
• Database design and implementation
• Custom Intelligent Search brokers for each of the content source websites
• Custom rules for Text Analysis entity extraction
• Database application for managing the supplier websites
• QA and UA testing
• Training (end user training; technical training)
• Periodic status reports

The exhibit below shows a screen shot of the top of a Company Profile page. It contains general company data, a vendor profile, the primary address, and 12-month purchase order commitments.

Screenshot of a Supplier Profile Page (Top)

The second exhibit shows additional information presented further down on the Company Profile page. This screen shot includes primary competitors and a complete list of human intelligence about the specific vendor.

Screenshot of a Supplier Profile Page (Bottom)

The third exhibit shows NAICS Codes, US and UK SIC Codes, and Hoovers Industry Codes for the vendor.

Screenshot of a Supplier Capabilities Page

Iknow first developed standard templates for describing the various types of proposed projects. The use of common templates helped describe each project consistently and make “apples-to-apples” comparisons between projects. Project owners were asked to describe their projects’ objectives, benefits, activities, deliverables, degree of technical effort, and timetable. Thirty-two projects were defined.

Iknow then classified the projects into three groups:

• Vision Shaping Projects. Vision Shaping Projects are those projects that establish a comprehensive, enterprise-wide direction for the company’s metadata architecture. These projects can be characterized as top-down and strategic.
• Foundational Projects. Foundational Projects are those projects that indirectly create business value by providing the infrastructure to support the company’s value-creating core business processes and activities. Foundational Projects address internal data and information management processes, functional architecture, content architecture, governance, and other information technology issues.
• Revenue Enhancement Projects. Revenue Enhancement Projects are those projects that generate new revenues or expand existing revenues, usually from new or improved products and services.

Iknow categorized the proposed projects into these three groups because the criteria used to evaluate the projects varies across the groups. We defined several metrics for evaluating projects; some of the top metrics included alignment with mission, goals, and strategy; alignment with customer expectations and/or requirements; cost (to implement the project); revenue generation; return of investment (ROI); and time sensitivity/urgency.

Iknow developed a nine-box scoring grid to visually demonstrate the relative position of the projects within each group with respect to the two criteria selected for the axes. An example of relative project ranking within a category is presented in the exhibit below. Projects highlighted in yellow (numbers 10, 17, and 18) have the highest relative ranking on the Customer Expectations dimension.

Project Evaluation and Selection Grid

A day-long workshop was held with senior business and technology executives to evaluate the potential projects and to select the projects for next year’s budget. The objectives of the workshop were to:

• Establish clear definitions for each of the potential projects.
• Propose any new projects.
• Agree on a common framework for project evaluation. This includes clear definitions of the evaluation criteria.
• Evaluate the projects against the criteria to determine their relative importance.
• Select the highest-rated projects for the roadmap.

Thirteen projects were ultimately selected by the workshop participants to be included in the final project portfolio.

Iknow conducted the following activities:

• Developed standard data collection templates and interview guides that were used for all interviews.
• Conducted individual and group interviews across all major functional areas with process owners, process participants, and end users of process outputs to identify and describe the knowledge assets that they used in their jobs.
• Prepared a detailed interview summary after each interview.
• Collected and reviewed the available standard operating procedures for the department’s key functional areas.
• Cataloged and stored the information about each knowledge asset in a custom Microsoft Access database.

The information was collected and aggregated into 15 categories that aligned with the Department’s major functional areas. The categories were:

• Clinical Operations
• Clinical Trial Management (Clinical Supplies)
• Customer Contact Center
• EBM and OR Strategy
• Learning and Professional Development
• Medical Affairs
• Medical Communications
• Medical Education
• Outcomes Research
• Planning and Contracts
• Quantitative Methodology
• Safety Operations
• Scientific Affairs Field
• Standard Operating Procedures (SOPs), Compliance and Auditing
• Trial Methodology

More than 80 Department personnel were involved in the Knowledge Audit and more than 700 individual knowledge assets were collected, described, and inventoried. Assets were categorized as either process inputs or outputs. Inputs were defined as knowledge assets that initiate, are accessed by, or referenced during a business process; outputs were defined as knowledge assets that are an end product from the business process.

Iknow approached the problem in three steps:

1. Current-State Assessment. Iknow comprehensively analyzed the firm’s environment from the perspectives of people, process, technology, and culture, and developed a current-state snapshot, or baseline, to characterize its operations. The team studied several core processes, including project delivery and business development. Iknow conducted interviews with the firm’s employees and reviewed the available tools, checklists, and other resources used by the consultants and support staff.
    
2. Recommendations. Iknow distilled the research and interviews and developed specific and actionable improvement recommendations. We also created a knowledge management strategic roadmap and implementation plan, complete with milestones and resource requirements.
    
3. Board of Directors Presentation and Approval. Iknow presented the findings from the baseline assessment and the improvement recommendations to the firm’s board of directors and other senior staff to gain approval for implementation.

One of Iknow’s recommendations was that the firm needed to invest the time and resources necessary for capturing, storing, organizing, and sharing the company’s existing knowledge and intellectual capital. Another important recommendation was to implement an enterprise search strategy. In the short term, Iknow’s search strategy included:

• Purchase and implement an enterprise search software tool. The search software would index the firm’s existing content and make it available across the organization. Restrictions and access safeguards would be implemented to protect sensitive and proprietary data.
   
• Begin a knowledge-capture effort. The purpose of this activity would be to identify and capture the firm’s highest-value knowledge assets. This activity would allow the firm’s consultants to submit knowledge at any time and from any location to designated knowledge managers.

The primary objectives of this project were:

• Design a comprehensive taxonomy and a set of WFLLC-specific content metadata tags.
• Store the new taxonomy and metatags in Smartlogic Semaphore. Semaphore is an enterprise semantic platform that augments traditional information management systems (such as search, content management systems, and business workflow engines) by adding advanced content classification, metadata, and navigation capabilities to deliver a more complete enterprise information management experience.
• Develop a new content repository in Microsoft SharePoint, a content management system (CMS).
• Migrate the WFLLC’s content from the current websites to SharePoint, remove old or dated content, assign the new metatags to the WFLLC’s content, and store both the articles and their associated metadata together in the new CMS.
• Configure the Coveo search engine to incorporate the new taxonomy, metadata, and content in the SharePoint CMS and implement the new functionality on the existing WFLLC websites.

Iknow performed the assignment in two major work streams: (1) Content Enrichment; and (2) Search Engine Enhancement. The key steps were:

• Content Enrichment Work Stream

1. Project Preparation and Initiation
2. Analyze the Existing Content and Perform Initial Processing
3. Develop the WFLLC Taxonomy and Metadata
4. Design the Content Tagging Process
5. Implement the Content Tagging Process

• Search Engine Enhancement Work Stream

1. Validate the Coveo Enterprise Search (CES) Installation
2. Design the CES Interfaces
3. Implement the New UI Designs and Search Tuning

Commercial software products from three product vendors were used to create the overall solution. The products and short descriptions of their functionality are provided below.

• Smartlogic Semaphore software was used for ontology model development and management, ontology-driven classification, and browsing search results. The specific software products purchased from Smartlogic were:
  • Ontology Manager—designed to allow multiple users to create, enhance, and browse all types of semantic models, whether they are lists, controlled vocabularies, taxonomies, thesauri, or ontologies. The software covers the lifecycle of taxonomy development and maintenance. The license includes unlimited semantic visualization (SV) web clients (i.e., unlimited use on the WFLLC’s websites).
  • Semaphore’s Semantic Enhancement Server (SES)—a high-speed XML-based index that allows developers to query an ontology or taxonomy in real-time and create and deliver topic maps, faceted search, visualization, topic pages, related content, and other user interface components.
  • Text Miner—used to automatically extract nouns, noun phrases, and other entity types from unstructured text content. The Advanced Language Pack provides entity extraction capability for a specific language. The English language was licensed for use by the WFLLC.
  • Content classification is the process of analyzing a document and adding metadata “tags” that describe that document. Metadata tags are sourced from a taxonomy or other form of controlled vocabulary. Modules of Semaphore Content Classification and Text Mining Server include:
    • Classification Server. The enterprise scalable classification and text analysis processing engine.
    • Rule and Template Editor. A client tool to generate the rule base templates and build custom rules.
    • Rulebase Generator. The processing stream that generates the rule bases from the Semaphore model.
  • Semaphore for SharePoint is a comprehensive integrated solution for Microsoft Office SharePoint Server 2007 or SharePoint 2010. This connector extends SharePoint by tightly integrating Semaphore’s automatic classification and taxonomy governance capabilities with SharePoint’s content management functionality. 
• Two Microsoft products were purchased and used in the Content Enrichment solution.
  • Microsoft SharePoint is a web application platform. SharePoint provides a central location for storing content such as files and documents. This content can be accessed and modified within a web browser or by using a client application (typically Microsoft Office) via desktop or smartphone. SharePoint 2010 provides a concurrent edit ability with Office 2010.
  • Microsoft SQL Server is a relational database server. Microsoft SQL Server’s primary function is to store and retrieve data as requested by other software applications, be it those on the same computer or those running on another computer across a network (including the Internet).
• Coveo Enterprise Search was used on the WFLLC websites. The Search Enhancement portion of this project improved the search functionality of the Coveo product. Two Coveo products were used in the Content Enrichment solution.
  • Coveo Enterprise Search, version 6.5. Coveo Enterprise Search is a modular and scalable enterprise search platform that indexes information stored in various repositories throughout the enterprise.
  • SharePoint Connector. Coveo’s SharePoint Connector is one of the best ways to integrate information stored in SharePoint with other information in Coveo’s search index. Coveo supports multiple SharePoint versions, including SharePoint 2010.

The solution was developed on Amazon Web Services and Server Intellect, a third-party hosting provider. The overall technical architecture of the Content Enrichment and Search Enhancement Project is illustrated in the exhibit below. The exhibit shows the three types of functionality that are integrated together in the overall solution:

• Taxonomy and Classification, provided by Smartlogic
• Content Management and User Interface, provided by Microsoft
• Search, provided by Coveo.

Content Enrichment/Search Enhancement Solution

The four arcs highlight the integration between these products:

1. The Semaphore taxonomy model and classification rules are used to tag the content in the SharePoint content repository.
2. The Coveo SharePoint connector accesses the SharePoint content repository during content indexing.
3. The Semaphore SharePoint connector provides enhanced search and browse functionality.
4. The Coveo SharePoint connector provides enhanced search and facet functionality.

The first phase of this effort focused on the data collection, data extraction, and data quality tasks. The approach involved four important steps.

• Target source identification and clean-up. Iknow worked with the company to update their list of source content, including removing inactive content sources and updating their content licenses.
   
• Build the connectors. Iknow built the individual connectors to enable the platform to collect information directly from more than 600 federal, state, and local government websites and indirectly from thousands of news articles published every day.
   
• Create customized rules. Iknow built more than 200 custom rules for extracting names and other entities and for validating the extracted entities. This customization is essential because certain name spellings, complex sentence structure, punctuation variations, and other language-related problems cause the software to miss or misinterpret the desired content.
   
• Load the outputs into the existing database. The extracted entities are converted into structured data and loaded into the company’s existing proprietary database.

Iknow designed and implemented the complete system. The SAP BusinessObjects Text Analysis software was chosen as the core processing engine for natural language processing and entity extraction functionality. The system also captures and stores the full-text articles to provide back-up support for every entry in the database. In addition, Iknow performed overall system quality assurance testing and validation, prepared and conducted end user and administrator training, and prepared the technical documentation.

Iknow conducted a current-state assessment that involved the entire 200-person marketing organization and developed an overall strategy for improving data and knowledge management.

The current-state assessment was conducted using a framework involving five core dimensions: strategy, content, processes, people and organization, and technology. Detailed interview guides were developed for two groups of internal stakeholders: customers of the marketing organization, and the marketing staff involved with the acquisition, analysis, and dissemination of marketing information. Twenty-seven interviews with key stakeholders were conducted and their input was used to develop a current state snapshot. Flowcharts of the current business processes were also prepared.

A comparison of current state with leading practices drawn from strategic marketing functions at other companies revealed significant improvement opportunities along all five dimensions. Iknow synthesized the interview findings, developed an overall strategy, and defined 12 projects as part of an overall roadmap that would improve the capabilities of the marketing organization.

Iknow designed the new knowledge management processes, identified sources of information and tracked how they were used, and created a performance scorecard.

The project involved three work streams. The Process work stream documented current state business processes and knowledge flows and identified key performances measures. Over fifty interviews led to understanding specific managed-care related tasks and activities, the decisions they make related to managed care, sources and uses of data, and “future state” business requirements.

The Content work stream focused on content identification and acquisition. A Content Source Analysis was conducted to identify the types and sources of information used by each of the groups. A database was created that cataloged all of the managed care content.

The Scorecard and Technology Architecture work stream focused on creating a measurement tool that provided visibility over major account status and to the performance of his internal operations. Iknow developed a performance management framework and defined the performance metrics. The technology architecture documented the company’s technical standards and identified the data sources required for the scorecard application.