Pre-Publication Version

Revised version published under the title "Lifting the Burden" in Information Management Journal July/August 2005  

Organizations beginning the process of implementing enterprise-wide electronic records management systems (ERMSs) must address the role the desktop end user should play in making records management decisions. Some organizations have adopted the practice that the end user decides whether a document or email is a record and assigns the document its place in the organization file plan.³ Other organizations attempt to remove as much records management decision making as possible from the desktop end user. Insofar as possible, these organizations cause their information systems to make records management decisions in the background, transparent to users, on the basis of workflow analysis and/or a set of pre-defined rules.

The question this paper addresses is: Which approach results in the highest quality records management and the most complete capture of records into organization recordkeeping systems at the lowest acceptable cost?

The authors answers the question by asserting the proposition that the highest quality, most accurate and complete records management is achieved within an enterprise when the desktop end user is minimally involved in records management decisions. In other words, records management of the highest quality and accuracy occurs when records management is non-intrusive to the desktop end user to the maximum extent possible, and does not interfere with the normal work routines of professional staff in the enterprise.

The proposition rests on the following arguments:

As a general principle, the introduction of new information technology (IT) to the workplace should increase, rather than decrease, worker productivity. Requiring desktop end users to make records management decisions lowers worker productivity because making the decisions takes time and attention away from the worker’s primary job. Therefore, it is undesirable as a matter of IT management policy to introduce new desktop procedures that intrude on the worker’s primary job and lower productivity.

Records management is the primary responsibility of the trained records management staff in an enterprise. Asking desktop end users to make records management decisions in effect makes every user a records manager. This is an undesirable condition for the enterprise. The cost to the organization of continuously training desktop end users to make accurate records management decisions, as well as the cost of implementing a quality assurance program, will be prohibitively high and represents an investment that enterprises cannot well afford.

Experience gained in those enterprises that require desktop end users to make records management decisions (e.g., using e-mail "pop-ups") indicates that the resulting quality of records management is very poor. End users know that records management is not their job and generally do not comply with the request that they perform records management functions. They "game" a system of email pop-ups so that almost nothing is a record or everything is a record. The result is that the enterprise achieves low quality records management.⁴

In counterbalance to these arguments, the authors recognize that end users are the principal creators and receivers of the bulk of business records. This fact makes it impossible and impractical to eliminate end users entirely from all records management decision making. Training end users to identify, and perform minimal categorization of, a record that will then be further examined and categorized by a staff of records management specialists can be more easily afforded and more effectively instituted. Identification and high-level categorization of records may be a minimum acceptable burden on the desktop end user. Hence, a practical alternative to no end user involvement is to ask desktop end users to make a minimum of high-level records management decisions if other non-intrusive techniques are not adequate to achieving high quality records management.⁵

Our principal data source for this article was four case studies we conducted with three U.S. federal agencies and one intergovernmental agency. The agencies were:

U.S. Government Accountability Office

U.S. Office of the Comptroller of the Currency, Department of Treasury

U.S. Nuclear Regulatory Commission

An intergovernmental agency that preferred not to be named

We chose these enterprises because previous knowledge led us to believe they would have materials to contribute to the study. The case descriptions below explain how each organization has attacked the problem of making records management non-intrusive to the desktop end user.

The U.S. Government Accountability Office recently purchased Hummingbird’s Enterprise integrated electronic document and records management system (EDMS/ERMS) for 3,000 seats. As of October 2004, the agency had completed acceptance testing and was rolling out pilot projects that covered the areas of:

Policy

Administrative program

Audit/engagement

The GAO Records Officer stated that she saw her task with respect to ERMS as "making records management as invisible as possible to the end user." The following items explain how GAO has attempted to make this happen.

Simplified File Plan. GAO has greatly simplified the agency records file plan. All records go into one of three "buckets": Mission/Engagement, Administration, or Policy. Each bucket has its own disposition schedule.

Subdivided Buckets. Each bucket is subdivided into a set of "functions." The total number of functions for all three buckets is 33.

User Profiles. GAO users are accustomed to filling out a document profile preparatory to saving a document in an EDMS. Saving to the ERMS will present the user with one new question box in the document profile. This box will ask the user to choose whether the document is mission/engagement-, administration-, or policy-related—one of the three buckets.

After choosing a bucket, the user must choose a function within the bucket. The user thus makes two "clicks" to describe a user profile in the ERMS—the first click is a bucket and the second click is a function within that bucket. Once a user makes the two clicks to describe a profile, all records management is taken care of while the user remains in that profile. The user may continue work on multiple documents, spreadsheets, etc., without making any more profiling clicks as long as the user remains under that profile. Given that the user has made the necessary clicks to define a profile, the records management functions for any document are accomplished when the user gives the document a title and hits "Save."

How much desktop end user decision making is involved in records management at GAO? At a maximum, a user would have to input two clicks and a document title to accomplish records management for any document. That is, the user establishes a profile/folder and titles the document at hand. Assuming the user is continuing to work within a previously established profile/folder, the user need only title the document and hit "Save"; records management is automatically carried out for the document.

The Office of the Controller of the Currency took possession of a new integrated EDMS/ERMS from Hummingbird Ltd. in October 2004. The agency had a phased implementation plan that called for concentrating on at-risk and mission-critical records in Large Bank Supervision (LBS).

As at GAO, the OCC Records Officer sought to make records management as non-intrusive as possible to users. The OCC simplified its file plan into big buckets and reduced its records schedules to 61 items. The major file categories constitute only 18 buckets. LBS uses the National Filing System, a national bank filing system, to file bank examination records.

Records management occurs in the OCC EDMS/ERMS in two ways: a rules-based method and a drag-and-drop method. In LBS, a major and critical part of the agency, the business processes employ a rules-based method of records management. When a bank examiner finalizes a bank examination report, the report will automatically be placed in the ERMS. OCC performed a workflow analysis of the bank examiner processes. At the end of the bank examiner workflow, the examiner or analyst fills in the date to move the document into the ERMS and at that point – and not until that point –a "Make Record" icon appears on the desktop. The user clicks the icon and all records management functions occur automatically. In this respect, records management is designed into the business processes.

In other cases, users will drag and drop files from the EDMS to the ERMS. For large-bank work, the user completes a document profile, required and enforced for all documents. In preparation for the advent of the EDMS/ERMS, OCC created a customized mapping table that links document profiles to the ERMS in the background. The user’s choice of Document Type and Name of Bank automatically maps to file locations in the ERMS.

OCC places a heavy stress on records management and staff members understand they must keep records of their work. Prior to the advent of the EDMS/ERMS, keeping records meant printing and saving the paper printout, a process that staff find onerous. The staff understands that until their compliance with the new EDMS/ERMS is excellent (as evaluated by the Records Officer), they must continue to print and save the paper printouts. Hence, they have a strong incentive to comply with the requirements of the new system.

The Nuclear Regulatory Commission's (NRC) integrated EDMS/ERMS is called ADAMS, which stands for Agency-wide Documents Access and Management System. ADAMS has been in operation since 2000. ADAMS originally employed the paired solution of FileNet's document manager and TruArc's ForeMost records manager; ADAMS now uses later generations of the same products as offered by FileNet and EMC Documentum. With a few exceptions, particularly in the administrative areas, the NRC employs only electronic records management practices. All agency records entering the NRC in paper format are scanned into electronic format for capture in ADAMS before being distributed.

NRC-generated documents undergo a two-step process to enter ADAMS.

Document Submission. NRC staff submit documents to ADAMS either electronically or in paper form. If in paper form, the staff complete a form that provides instructions for scanning the document into electronic format. When submitting a document to ADAMS, the staff complete a minimal profile of the document. The minimal profile supplies the following information:

· Brief title

· Public availability

· Sensitivity and access rights if the document is sensitive

· Template number

· Any unique identifier not visible on the document that staff subsequently may use to retrieve the document

Document Processing Center (DPC). The submitted documents go to the DPC. DPC is an on-site contractor-operated facility employing approximately 40 fulltime personnel. DPC performs the following activities:

· Completes and ensures quality and accuracy of document profiles

· Scans all paper pages and conducts 100 percent image quality control

· For packages, enters paper documents and electronic files in the sequence specified by staff on the form

· Declares documents as official agency records

· Converts electronic documents submitted to DPC into PDF, ensuring integrity of converted files

· Copies records to the Publicly Available Records Systems, the public access version of ADAMS, in accordance with the date specified in the "Date to be Released" property

· Audits, monitors, and assesses problems and trends

NRC staff perform only limited records management functions. DPC takes the actions to declare a document a record and, along with select NRC records management staff, to assign it to a particular record collection, in accordance with the established files maintenance and disposition plan. NRC also has a records audit program that functions as a quality control program as well. The offices' record collections are audited routinely to ensure that documents designated for retention as records in ADAMS are in fact being placed in ADAMS.

We interviewed an intergovernmental agency that preferred not to be named in this article. The agency is a complex, diverse, and highly structured intergovernmental organization working throughout the world in a broad range of development areas.

The agency described a vision for implementing records retention and disposition at the document object level within a broad range of electronic content. The challenge was twofold:

To prevent records born digital in a multi-function environment (creation, publishing, learning, collaboration) from disappearing for lack of a records determination, and

To prevent overload of electronic content systems’ capacity resulting from inability to dispose of non-records content (copies of convenience).

The goal behind this approach was to achieve a better balance between over- and under- retention or disposition of content.

The vision was based on the need to identify and distinguish records and copies of convenience. The vision assumed a rationalized retention and disposition schedule exists at the enterprise level. Further, the vision assumed that document templates used to create content could include the metadata required to identify records and copies of convenience, to associate schedule information with documents born digital, and to predict and associate the disposition or retention date.

The templates can provide the information needed to perform a best guess at predicting retention and disposition actions. Retention and disposition actions will not be performed automatically within a software system. Rather, the predicted appropriate schedules and retention or disposition dates will be aggregated into reports to records managers and systems owners who will then carry out the actions. The system would generate periodic reports that will be the basis for the records manager/system owner to take aggregate scheduling, retention, and/or disposition actions.

For example, a document template used to create news stories would contain retention and disposition-relevant metadata, such as:

Date the news story was created

Date it was published,

Creating organizational unit

Creating author’s corporate affiliation,

Business process with which story is associated,

Where document resides (e.g., website versus in an agency database),

Version (e.g. final or draft),

Tagged by the author as a record or convenience copy

This substantial amount of metadata would permit a best guess as to whether a retention or disposition action should be taken, and the appropriate retention or disposition date. With the metadata embedded in the document template, the information could be checked before any deletion action was taken (reducing loss of records) or used to advise that a retention action is needed (increasing retention of records).

In this approach users do not have to make conscious retention or disposition decisions. The benefit to records managers and systems owners is that periodic deletion of older copies of convenience for system capacity purposes can be managed without increasing the probability of loss of records.

The agency has already developed the metadata and reference sources for three kinds of document templates: Content Types, Business Activity, and Organizational Unit.  The agency has created eight content types, which are further divided into sub-content types.  Each sub-content type is associated with a document templates. The document templates would carry basic and common metadata elements required for retention and disposition actions and would also have a logical link to retention and disposition schedules.  

At this writing, templates have been identified, but the logical link with the records schedules has not yet been made.  Once the vision is fully realized, the agency would have the capability to push records management "upstream" into document management, merging records creation and management with document creation and management.

The planning and care taken in designing the user interface with an EDMS/ERMS are critical to the quality of records management in any enterprise. While our study was based on research in governmental organizations, we believe its findings are generalizable to any complex enterprise. Following is a summary of the strategies and tactics uncovered in this study.

Simplifying File Plans. These enterprises greatly simplified their records file plans. Simplifying file plans meant that the choices and decisions to be made in the ERMS were reduced. Big bucket file plans are in the nature of groundwork for non-intrusive ERMS implementation and have other appeals and uses beyond those discussed here.

Records Managers and IT Managers Working Hand-in-Glove. In the enterprises studied, interviewees consistently stated that the non-intrusive measures were successful because, from the outset, the records managers and the IT managers worked closely together. They reported that, once records managers and IT managers understood one another’s perspectives, they were able to jointly conceive innovative techniques and solutions that neither would have invented independently. In only one case did an enterprise report difficulties because of inability of records managers and IT managers to understand one another’s point of view.

Use of Document Templates with Embedded Records Schedules. The intergovernmental agency’s advocacy of document templates for conducting its work is probably the least intrusive strategy encountered in this study. The array of document templates developed by the agency reflects detailed workflow analysis of how the agency accomplishes its work. The user will never be presented with a records management decision as such, only with a document template decision. This makes records management completely non-intrusive at the desktop level. The downside of this approach is the significant investment the agency has made in developing and enforcing a comprehensive document template regimen.

Centralized Document and Records Management Processing. While centralization is not necessarily a virtue in itself, NRC made centralized document/records processing work for greater quality and cost control, as well as non-intrusion. Enterprises might choose to centralize within lines of business and the tactic may work just as well.

Use of Workflow Analysis. Two enterprises studied performed workflow studies to discover where in the chain of work activities records management decisions were made. They then focused ERMS implementation at those particular points in the workflow. Understanding a line of business or work unit’s workflow will always be a key to simplifying ERMS implementation and making it non-intrusive to desktop users.

Simplification by Work Unit. GAO divided its simplified file plan into three buckets by work units. A workflow analysis showed that GAO staff working in administration work in that one bucket only. GAO designed implementation of EDMS/ERMS for administrative staff so that only a single bucket, Administrative, appears on the desktop of personnel in the administrative area. The slogan at GAO is "two clicks and a document title" to accomplish records management. In the case of administrative personnel, one mouse click and a title is sufficient.

Persons with experience and expertise in the implementation of EDMS/ERMSs in governmental enterprises—and the numbers of such persons were not large in late 2004—appeared unanimous in the conviction that records management is a "back office" function that should not intrude upon the desktop end user. This conviction occurs most strongly among governmental personnel who have actually implemented an EDMS/ERMS and among industry personnel who sell to and service the government ECM marketplace. Collectively, this group is convinced that minimal intrusion upon desktop end users in records management decision making is a best practice in implementing EDMS/ERMS.

----------------------------

Footnotes: