Maximizing Your Data Repository Searches: Coding Choices

In the "Paper Chase" article, we discussed the basic elements of document management and touched on indexing and coding. It is through your indexing and coding system that your document repository turns into a useful and powerful tool.

Scanning and coding documents have obvious cost savings over handling paper. With paper, you encounter repeated costs each time you need to copy the document set — for trial team members, consultants, opposing counsel, and so forth. Add to that the labor costs of keeping each set updated as new documents are added… well, your costs skyrocket quickly.

In terms of cost, you have several choices to make as you contemplate a digital repository. Simply scanning your documents into images is only a slight improvement over paper. You don't have to lug around boxes, but you don't yet have an efficient way to access and search those files.

To make your file set manageable, you need quickly to understand what's in each file and be able to jump immediately to the most valuable information. Each file needs an electronic "label" or "code" so the computer can sort and organize the library. Coding is organized into three levels, from basic to complex.

Coding Level Choices

Level 1 Coding In this basic coding work, we can organize the digital files by referencing "load" files — basic instructions about a document that assist in organizing and sorting scanned images. We can attach load files to document sets for you. Without these load files, only the last person to work with a file will know where it's stored, and the file can't be included in a searchable database. (Ever try to find a file on someone else's computer — next to impossible isn't it?) In Level 1 Coding, we simply fill in some basic fields about each document such as document date, type, author, recipient, or other simple identifiers. These fields are how you search through the database. The benefit to you is obvious. You could ask for all of the documents from a specific author, on or around a specific date, to a specific recipient. Viola! All documents meeting those criteria are available for you to se, by selecting from a list of docs presented. You can review each one, or further refine your selections to find the exact document you want. Level 2 Coding In this level, we're now linking a more elaborate set of key words to your scanned images. You need highly experienced and trained coders who can review every single document and link them to helpful and intuitive key terms, issues, mentions and phrases. We work directly with our clients at the start of a project to set the criteria and standards for search terms. Only those manually entered key words and phrases are available later as search terms. The additional coding makes your database even more valuable. Searching in a level 2 database will retrieve every instance of the term or phrase, and then take you to its context. Gone are the days of sending a paralegal to the warehouse for hours or days to pull out every piece of paper with a particular chemical name! Level 3 Coding Level 3 Coding creates the most complex and complete database search capabilities. As your documents are scanned, they are also processed by optical character recognition (OCR) software, which identifies each number and letter shape on the page (as opposed to creating a static image of the page). The OCR program converts this data into a text file so that the search engine can recognize each word. Because OCR is seeking letter and number shapes, the cleaner and crisper your originals the more accurate the text file. If your documents are blurry, third- or fourth-generation photocopies or faxes, then the OCR accuracy rate begins to drop somewhat. The benefit here is complete mastery over all content, associations, references and relationships of all people, places and descriptions in your database. The likelihood of discovering the trends you suspect are extremely positive, if the story can be told in the documents you have in the database.

Documents versus Pages