I wouldn’t recommend either of those approaches (although yours is closer to intelligence than your IT dep’ts!). What I would suggest would be a multi-vectored approach:
Identify known high value/high risk (HV/HR) and vital records manually as these are usually the ones easiest to find and classify manually. This can be done using tried and true content inventory practices (usually Excel based as you mentioned)
As part of the content inventory identify all locations (logical electronic and physical) that have content/records/data that might be desirable to migrate
Acquire and deploy a proper discovery and auto-classification solution and point it at the repositories identified in step 2. The specific solution is dependant on what your EDRMS system is (I assume it is OT?). Once the content/data is appropriately tagged with metadata actual migration becomes a quite straightforward mechanical process which can be handled through whatever mechanisms the EDRMS supports (be that scripting like PowerShell, Windows scripts, or actual ingestion tooling for the platform)
Obviously the taxonomies (plural not singular) you create in the target system depends to a degree on what your discovery process finds but, for any large organization like BCPS, typically end up being:
Controlled – this is the content/data that sits somewhere middle and above on the 2 axes of “value” and “risk” and would fall under your retention schedule for both compliance and business retention reasons. Obviously much of this will also have sensitivity classifications that need to be applied along with permissions levels and such. Ideally, your EDRMS should support at least 2 ‘structures’ for your taxonomies; one for the retention schedule classification and a separate one for how the ‘business’ thinks of and interacts with the data. For the second structure I’m personally a HUGE proponent of “Functional Decomposition” which maps the content/data to your business FUNCTIONS starting at a very coarse-grained aggregation and working through multiple levels of decomposition into finer-grained. As a general rule, if you begin at the “Capabilities” level (which means developing a Capabilities Map if BCPS doesn’t have one), then elaborating that to the macro-level Functional Map with further elaboration dependant on each stream within the top level map or not being based upon the benefit realization value of doing so (different functions and sub functions will have differing levels of benefit realization value that can be achieved so the effort should be managed accordingly). As you decompose you get to the point of the ‘Tasks’ that make up the Functions and that is where the taxonomies stop. A particular sub-category of content/data that should ideally be identified here is BR/DR (Business Resilience/Disaster Recovery which also obviously includes Vital data), as every organization SHOULD have learned this year. Here is a link to a graphic on my site that illustrates the WHOLE process if you have mandate and budget to do it right from start to finish: bit.ly/36NX1mV
Departmental – this is the content/data that isn’t really aligned to the Functions the organization executes in order to deliver the Capabilities. Rather it is content/data that really only relates to a specific departmental context and has very little to no cross-department aspect. Hence, this is always a fairly small percentage bucket
COP (Communities of Practice) – this would be content that is produced by any specific communities that may exist within the organization. An example within every large org would be the “social committee” type entity. But others may exist. To stress though, this is NOT content/data related to the orgs Capabilities, Functions, or Tasks per se
Reference/External – obviously any content/data from outside sources that must or should be maintained but is not produced by any internal source (staff. Contractor, or retained consultants)
Email – I break this out as it is ALWAYS the most difficult to deal with since a single email can, and very often does, contain information, either in the body, attachments or both, that falls under multiple retention buckets. Even though the discovery/autocategorization tools have become much more capable over the last few years, this still presents a bear of a problem. I generally advise doing all of the OTHER repositories and getting really good with the autocategorization tool, the IA, governance, etc. and THEN applying that to email.
Hope that helps.
Vice-Chair, ISO Trustworthy Content/Document Management Working Group (WG)
Standards Council of Canada (SCC) ISO/TC 171/SC 2 Mirror Committee (MC) Member
Sent: 11/17/2020 12:13:00 PM
From: Colleen Loguisto
Subject: Migrating from Shared drives to EDRMS system – Manual or Auto
We need to migrate millions of records from our shared drives to EDRMS. I am being told by our IT that we need to migrate our files manually. I was shocked to hear this. Apparently it will take someone about 50 years to migrate everything. Crazy life in records. LOL. I am wondering if our IT people simply haven’t caught up to electronic records management processes or if this is how it really is.
I thought they would follow a 2-step process where they would 1. auto-generate the folders in EDRMS using a detailed excel spreadsheet and then 2. auto-migrate the files using the same or another spreadsheet which mapped specific shared drive folder to the corresponding EDRMS folder.
Has anyone ever auto-migrated files from shared drives to an EDRMS system? Any ideas would be helpful. TIA. 🙂