Data exception reports that result from incomplete processing are a core requirement of any eDiscovery data processing system. No system can successfully process every file that has been gathered. Among the top reasons for processing failure include: unsupported file types, password protected files, unknown file types, files without text (such as pictures, gifs, etc) or corrupt files. Generally, most systems output a spreadsheet of all the files that did not process and it is then up to a series of project staff to painstakingly locate the original files, examine them and determine whether further resources should be expended on revealing their contents. It is important for any service bureau to handle processing exceptions quickly and cost effectively to provide clients with the most complete processed data that is possible. What can be done to make this important process less time consuming and more productive?
Improving on the utility of exception reports is one way to shorten the entire exception process and assure that the needed files are promoted to the database and that the irrelevant files are set aside, removing them for the need to be reviewed. We determined that having interactive exception reports that are directly connected to the database makes the reports infinitely more user friendly. Since Cavo eD is a single unified product with one database, creating user friendly interactive exception reports was the obvious answer.
Cavo eD exception reports are a series of graphic representations with three major headings: Processing Results, Exceptions by Categories and Exceptions by Type. In this first graphic, under each of the major headings a pie chart is presented with the categories, percentages and number of files clearly indicated. Additionally, the document grid appears below the graphs and will display whichever of the categories is clicked on by the user. Individual documents can be opened directly from their location in the preprocessed file and additionally, we include the original media path so the document can be recovered directly from the original media. We also track the complete pathname to the file in question so that it can be examined in its preprocessed state for a quick determination of whether it should be ignored or further examined.
The second graphic that is displayed represents what happens when the user clicks on one of the Exceptions by Categories to focus attention on specific documents. In this instance, the user clicked on the pink portion (fatal exceptions) which number 14,740. Fatal exceptions are files that cannot be promoted directly to the grid because they could not be processed on any level. Note that the grid below is now showing all 14,740 of these documents. The user can then begin to filter and sort the list and determine which files can be ignored (filtering on .dat files yields 11,534 files can be immediately ignored). Other files can then be examined by clicking on the document link to determine if they should be resubmitted for processing.
The third pie chart, Exceptions by Type represents the Non-fatal exceptions portion of the second chart, which means that the document did not process completely but might be directly promotable after a quick human review. (Some non-fatal errors like Password protected files cannot be automatically promoted and must be further investigated). This chart displays the major categories of exceptions, showing why documents were not completely processed. Certain categories of documents can be promoted directly to the Grid once a determination has been made by a user that there is enough information about the file to make it immediately useful to the team. In the instance portrayed in this third graphic, the user has clicked on the OCR Disabled category, and it shows that 14,574 files appear in the grid for examination. When the user filters on .jpg, the grid shows 8,216 documents. After examining a few of these to confirm what they are, the user can promote these directly to the database by clicking on the “Promote to the Grid” Button that appears in the Title Bar which will promote these documents to the review file without the need for reprocessing.
All the exception reports can also be exported to a .CSV file for further review using existing procedures. By adding interaction to the exception reports, users can more quickly access documents and make determinations about their status, and immediately promote those documents that they determine should be promoted.
The Way eDiscovery Should be Done…