Let me clarify what I mean by the hosted e-discovery model, because one size most emphatically does not fit all. At the low end you have a Storage Service Provider (SSP) that stores customer data and grants remote access to review and analysis software. This is not a terribly common model, since most corporations that are subject to frequent litigation are unlikely to entrust their data to a small and unspecialized storage service. I am not referring to those services.
At the higher end, the hosting service is dedicated to offering e-discovery products and services on hosted customer data. Most of these vendors are rooted in review and analysis, since this stage is the most extensive and the most expensive of all the e-discovery processes. In addition to the high ROI of review and analysis, other differentiation examples include highly optimized storage and high-performance servers for high-speed ingestion and processing. Some vendors have internal cloud structures, which allow them to take advantage of dense computing infrastructures. And all should offer high levels of security and user/role-based access control, with the corporate customer deciding who gets to see what. (Picture the difference between a first-year contract attorney and a 20-year trial veteran in terms of sensitive views.)
To be successful, the e-discovery host must maintain a set of strict requirements. The model must: a) maintain the highest security around the physical plant, user access, network, and data protection; b) offer high ingestion and processing performance with near-perfect uptime; and c) provide top-level review tools including multi-tier reviewing, sophisticated query and results support, excellent analytics, and early case assessment ability. There are other offerings as well that increase ROI and provide competitive advantage, such as multi-matter management, workflow management, process analytics for cost tracking, internal consultants, clouds, etc.
Clearly not all corporations use the hosted model. When corporations first look at e-discovery products for their internal use, they often start with a review and analysis package for their in-house attorneys (who may require their outside law firms to use the same product). Very often the same corporation will take the subsequent step of purchasing collection-based software to help with the massive job of classifying enterprise data from multiple data sources. (Thanks to growing data stores, this is a big job that is getting bigger by the minute.) This software helps them to classify and collect electronically stored information (ESI), to establish litigation holds, to get a handle on the merits of a case and to reduce the size of the data store for review. At that point, they must process or load the relevant data to their platform for review and analysis. Depending on the scalability of their review and analysis solution and their storage infrastructure, they may choose to do this in-house, send it out to their attorneys to do, or employ a specialized e-discovery hosting provider.