Enterprise search is the organized retrieval of structured and unstructured data within an organization. Properly implemented, enterprise search creates an easily navigated interface for entering, categorizing and retrieving data securely, in compliance with security and data retention regulations.
The quality of enterprise search results is reliant upon the description of the data by the metadata. Effective metadata for a presentation, for example, should describe what the presentation contains, who it was presented to, and what it might be useful for. Given the right metadata a user should be able to find the presentation through search using relevant keywords.
There are a number of kinds of enterprise search including local installations, hosted versions, and search appliances, sometimes called “search in a box.” Each has relative advantages and disadvantages. Local installations allow customization but require that an organization has the financial or personnel resources to continually maintain and upgrade the investment. Hosted search outsources those functions but requires considerable trust and reliance on an external vendor. Search appliances and cloud search, the least expensive options, may offer no customization at all.
Enterprise search software has increasingly turned to a faceted approach. Faceted search allows all of the data in a system to be reduced to a series of drop down menus, each narrowing down the total number of results, which allows users to narrow a search to gradually finer and finer criteria. The faceted approach improves upon the keyword search many users might think of (the Google model) and the structured browse model (the early Yahoo model). In the case of keyword search, if the end user doesn't enter the correct keyword or if records weren’t added in a way that considers what end users might be looking for, a searcher may struggle to find the data. Similarly, in a browsing model, unless the taxonomies created by the catalogers of an enterprise's information make intuitive sense to an end user, ferreting out the required data will be a challenge.
Enterprise search is complex. Issues of security, compliance and data classification can generally only be addressed by a trained knowledge retrieval expert. That complexity is further complicated by the complexity of an enterprise itself, with the potential for multiple offices, systems, content types, time zones, data pools and so on. Tying all of those systems together in a way that enables useful information retrieval requires careful preparation and forethought.