Pharmaceuticals

Business requirements: As proprietary internal data repositories continue to grow exponentially, pharmaceutical companies' decision makers are often overwhelmed with excessive aggregate data. The net effect is that they're not always able to find right information at the right time. In order to be a successful player, a pharma company needs to improve agility, the quality of its insights, and make quicker go/no-go decisions. Current industry wisdom indicates that every day a big pharma company can save in bringing a product to market represents $1 million in savings. Big Pharma clients are looking for search solutions that interface with their diverse data sources, collaborative document repositories, intranets, wikis, document management systems and shared file storage. The ability to share the appropriate data with high value employees in one place, and to enable social networking based on that data is the holy grail. Security and cross-authentication is also a major concern.

ESR consulting has been working with FAST and a major pharmaceutical company to create an architecture that provides a common point of enterprise connectivity through search technology and security. The client requires reliable indexing and information presentation of large data sources, as well as secure authentication and appropriate transparency on the user side. To achieve these and other business objectives, firms recognize that it is prudent to use a phased approach to enhancing its searching of information. The initial solution must be able to provide tangible benefits for everyday document search. This is not trivial. It involves creating a tool that ties together numerous information sources, both internal and external. From there, the initiative should then provide for more advanced solutions that take into consideration each business units goals and primary business drivers.

Integration points: The rapid growth and expansion of enterprise data sources on legacy and newly evolving systems make it necessary to rapidly evolve massive data connectors, while at the same time controlling the hardware footprint of the search system and stewarding the enterprise's data center resources. Integration with today's enterprise requires reliable and fast connections to such industry standards as for example Oracle, Lotus Notes, Documentum, Microsoft products, intranets, file shares and other points of specialize data integration. Security concerns require that there is appropriate cross-authentication impacting the desired transparency of the user's search experience.

Time to market: ESR and FAST are working with the client to design a system that can deliver enterprise reliable search in a phased approach, bringing the enterprise to search capability in a rapid and scalable fashion. Since information is a measurable currency in this industry, there is a positive bottom line impact from rapid deployment of each phase. This project is a model for both enterprise scalability and rapid testing/deployment planning.

Intranet Search

Intranets pose their own search challenges, especially when dealing with financial institutions, and others with highly confidential business information. ESR has deployed intranet search at several Fortune100 financial institutions, addressing such requirements as:

  • Integrated LDAP and scoped search
  • Forbidden query terms
  • Connecting to disparate content systems
  • Adopting operating standards of the enterprise
  • Preventing unwanted indexing of secure content
      • Content discovery:
      - Understanding who is publishing content
      - Determining where content lives
      - Monitoring versions of content
      - Understanding relationships between disparate content

Publishing

Having helped scientific and other journal publications bring search to their content stores, ESR is familiar with the unique changes faced in this sector. Some of the common issue addressed with media publishers include:

  • Content de-duplication
  • Continual updates and indexing
  • Providing tools for administrators to remove content from the index on demand
  • Tight integration with publishing content life-cycle processes
  • Large data through-put, and continual updates to indexes
  • Multi-lingual search and content processing
  • Subscription models and scoped search results

Yellow Pages

With extensive experience in deploying business directories, ESR has helped launch several major directory-search installations. Leveraging technologies such as geo-search, and “sounds-like”, these installs handle high query volumes and deliver quality results while leveraging complex business rules to support their revenue model. The challenges of Yellow Page type searches include:

  • Tens of millions unique records
  • Timely updates and removal of records
  • Complex business rules determining rank and visibility of results
  • High volume query management
  • Data cleansing and massaging to expand on and clean-up generic business data

Vertical Search

Providing a specialized search engine for niche verticals, ESR helps create sites where the audience is more focused, and the traffic of higher value to advertisers. Working with FAST, ESR has created verticals that are language specific, country specific, and domain specific. Leveraging our in-house RSS connector framework for greater throughput, these installations maintain large lists of URL feeds, use custom configured crawlers, and also leverage customized connectors to internal repositories ultimately creating well-defined, highly specific content, tailored for niche audiences and the long tails of the web.

Scientific Research

At a center for world-class scientific research, ESR was tasked with integrating FAST into an established search environment. One of the requirements of the customer was that we maintain functionality of their current query language so as to not have to re-train thousands of longtime users on a new tool, and allow them to continue their work unencumbered. To do so, a query translator was constructed from ANTLR, an Open Source tool used for grammar recognition. The tool leverages a template style approach which allows customization on the fly, should either the source or target language update.