“A force multiplier refers to a factor that dramatically
increases (hence ‘multiplies’) the effectiveness of an item or group.”
The biggest challenge when migrating content to an enterprise content management system is leveraging or multiplying the effectiveness of domain experts who determine which documents to preserve as records, which retention schedules to apply, and which data elements to extract from the documents. Without such leverage, migration efforts can require dozens of people working for a year or more. With so many people involved, many will not really be domain experts, and consistency will be almost impossible to achieve.
When domain experts have to review multiple copies of the same document types, the process becomes hopelessly bogged down and inconsistent decisions creep in. In some systems inordinate amounts of time are spent writing and testing rules to identify certain classes of documents, and only after all that work is done can the domain expert apply retention schedules and extract data.
With visual classification, documents are placed in groups or clusters automatically without requiring the time or attention of the domain experts – the software does the upfront heavy lifting. The domain expert can start his or her work with an awareness of what the document clusters look like and how many documents are in each cluster. Millions of documents will typically resolve into thousands of clusters and the domain expert can review exemplars from each cluster to make retention decisions and, using visual coding, can specify which data elements to extract from each cluster. A small team of a domain expert and three assistants can review 1000 clusters per day and perform visual coding on 300 of those clusters, averaging 9 or more fields per cluster. In a matter of weeks or a few months, the team can review and classify millions of documents. The small size of the team also makes it possible for the domain expert to provide 100% QC for the review process.
Visual classification has other advantages:
- Non-textual documents like maps, diagrams, plats, schematics and image-only scanned documents are also clustered visually, making it possible to accurately classify a far higher proportion of records than text-only systems.
- The system provides alerts when new document clusters are formed, making for up-to-date classifications.
With visual classification, large-scale migration to ECM’s or consistency audits of current ECM content becomes completely manageable whether that ECM is Documentum, FileNet, or OpenText. It is also an opportunity to capture the domain knowledge of experienced document managers and put it to use for the corporation for years to come.