Consider the various aspects of document management: Web-based storage vs. in-house servers; simple index vs. advanced taxonomy with metadata; all-in-one systems vs. interfaced open systems; searching time vs. indexing time.

In all of these comparisons one thing emerges as the single most important factor; the validity of the index assigned to the document images. 85% correct isn't adequate, nor is 90%, or even 98%. An index must be 100% correct 100% of the time, or there will be lost images. These losses translate into to wasted time because you will not know that you cannot locate a document until after you have spent hours searching for it.

Think of the time it takes to locate a paper document that is missing. Searching for an image can be even more time-consuming. Therefore, indexing correctly is an essential part of document management.

Indexes can consist of a few key fields per document or the full text. There are several different ways to capture the information that goes into the index. Many companies use automated character recognition on a number of levels from barcode to ICR. Others use key-from-image operators. Whether the initial process of gathering information is automated or manual, document management systems need a quality assurance process that guarantees accurate indexes. The bottom line for any type of data capture is that the system must have a way of validating the information through a second operator, a database lookup, or both, if you want a useful, accurate index.