More and more institutions are working to reduce the number of categories in their banking records retention schedules. This process results in what is commonly called a “big bucket” retention schedule. It should come as no surprise that many institutions are making these types of reductions. Customers are going digital and paperless quickly, and the insurance and financial services sectors are following suit. During this transition, many firms are being held back by their antiquated banking records retention schedule.
When records were largely paper-based, fine-grained retention schedules and records storage schemes were needed to find them. In contrast, when records are created and stored electronically and can be searched using full text and metadata, having too many retention classes makes it harder to automate the classification and scheduled disposition of records. This structural mismatch between digital systems and paper-based methods is one of the key reasons why many information resource leaders in the financial services and insurance industries are working to reduce the number of categories in their firms’ banking records retention schedule.
Too Many Choices
In a 2007 Iron Mountain-Cohasset Associates survey of big bucket practices practitioners found that fewer retention categories resulted in higher classification accuracy and an increase in scheduled destruction, regardless of whether classification was done manually or by using software. These results are also reflected in studies conducted by scholarly researchers. For example, in a 2010 paper titled, “What Does Classifying More Than 10,000 Image Categories Tell Us?” the authors found that as the number of categories in a data set increases, the accuracy of classification algorithms decreases.
Whether using their intuition or data-based science, more information managers have decided to reduce the size of their banking records retention schedule. According to a recent benchmarking survey of financial services and insurance firms, some 32% of financial services firms and 20% of insurance firms reported they have had fewer than 100 categories in their records retention schedule. Among the same respondents, 44% of financial services firms and 43% of insurance firms claimed to want fewer than 100 categories. Across these survey results, a majority (more than 70%) of respondents reported they wanted between 25 and 249 categories.
Big Bucket Best Practices
This survey shows many institutions need to reduce the number of categories in their schedule to reach their desired levels. Many institutions are going through the process of globalizing their retention schedules, which can be simplified by reducing the number of categories.
Even for large banking records retention schedules, most of the bulk is usually in the subcategories of record classes, not in the parent classes. It is not unusual for a 250-category retention schedule to have only 25 main categories and 225 subcategories. One easy way to start to reduce the number of categories in this schedule is to combine all subclasses within a main category that has the same retention period. In cases when the copy of a record or a subclass of records — e.g., “mortgage origination files” — can all be found in a single system, such as a mortgage origination system, classification and destruction can be simplified by keeping that subclass together and pointing directly to it while designating all other mortgage origination documents as non-records.
It is also true that many subcategories are derived not from a legal requirement, but from a business need. When business users claim to need a separate subcategory, they should be asked to provide examples of records from that class so that algorithm-based classifiers and people can be trained against the examples.
It is also important to ensure each category is unique. If multiple categories contain “IRS Form 10-4,” no person or algorithm will be able to classify them. In this case, employee 10-4 forms should be rolled up into the employee’s record and the 10-4 forms of the mortgage applicants should go into the mortgage origination file.
By using best big bucket best practices, organizations can make it easier to find the documents they need and know when it is time for records to be saved or destroyed.