Glossary

Conceptual Search - Technology that enables you to search for the concept or meaning of what you are interested in. This extends your reach into information that is otherwise inaccessible with traditional keyword search.

Clustering - Technique used to group documents that satisfy a set of common properties. The aim is to group related documents together to investigate prevailing relationships or overarching topics. Clustering can also be used to expand a user query with new and related index terms (suggest related terms and refine a query further).

Parametric Search - A type of search that provides a user with a number of parameters and range of values to use as criteria. For example, a realtor site may have a search that allows a user to narrow search results by price ($300K to $600K) and number of bedrooms (3 bedrooms and 2 baths). A typical difficulty for users with this type of search is that they will over-constrain the search and get no results, so advanced search implementations will also find close matches by relaxing some of the constraints.

Scoped Search - A search that applies to only a subset of the entire content, in order to narrow down results. Scoped search should be used carefully because users will often choose the wrong scope and get a worse match than when searching the entire database. However, when security content is critical, scoped search is very useful in segmenting secure sets of documents.

Near De-Duplication - Techniques that identify and associate documents that have mostly the same or highly similar content. For example and email reply with a one line “Thanks” is a near duplicate of the original mail. Near de-duplication often involves statistical analysis of document contents to determine degree of similarity.

Enterprise Search

It is estimated that as much as 80% of all information in the enterprise is unstructured or semi-structured data living in emails, text, voice and video. Enterprise search is no longer a simple convenience, it is critical to your company’s continued success.

To complicate matters, the explosive growth of unstructured information in the enterprise has been accompanied by a steady stream and often confusing array of search technology and products grouped together under the “Enterprise Search” banner. Understanding these technologies and making the right choices is not easy. Making wrong or inappropriate choices can be very costly and can lead to lost opportunities, revenue or competitive standing.

One Size Does Not Fit All – Information Retrieval Strategies that Meet Your Business Needs

Today, your users are demanding more than the simplicity of single text box to find what they need:

Sales departments want to view search results by market segment.
Pharmaceutical researchers want more powerful Conceptual Search capabilities to “mine” medical literature to discover new gene therapies.
Customers need to quickly find products by price, brand and other features.
Legal professionals need to quickly understand the major themes or concepts in a large collection of emails to prepare for a case.

Each of these scenarios requires a carefully planned information retrieval strategy to serve the needs of the user and to find the hidden opportunities for your business. Lawyers and research scientists may need strong clustering and visualization capabilities to understand important concepts in large data sets. Retail customers may require parametric search capabilities to refine results to products they are most interested in.

Our process starts with understanding your process. We evaluate your information and how it is organized, work with you to define information retrieval goals and most importantly, we evaluate the relevance, quality and value of the results.

The Cost of Getting It Wrong

In legal electronic discovery (e-discovery) finding, reviewing and analyzing the right information can mean the difference between winning and losing a case.

The discovery process poses many challenges to organizations:

Data sets in multi-party litigation can be very large (millions of documents).
Review and culling time is expensive and error prone, especially in multi-party litigation.
For international litigation, content may be written in multiple languages.
Boolean and keyword search are inadequate for finding documents that do not contain the search terms but are still relevant.
Security in multi-party litigation is paramount.

E-Discovery demands a more holistic approach to information retrieval:

De-duplication and near de-duplication technologies may be required to minimize the number of documents that need to be reviewed.
Cross lingual search capabilities may be required to search content in multiple languages.
Conceptual search technologies may be required to assist user in formulating queries and finding documents that contains same meaning but different keywords.
Scoped search capabilities may be required to secure a group of documents.

Your Partner in Success

Taking a broader approach to Enterprise Search as part of an overall information retrieval strategy is critical to your success. Enterprise Search is an essential component of your business process and a solution that requires attention to the unique needs of your business. Digital can help you make the right decisions about Enterprise Search, support you from concept to rollout and make sure your projects are a success.