To paraphrase Patrick Lambe (yet again):
Building classification schemes is at the same time deceptively simple and fiendishly complicated.
It’s deceptively simple because it’s one of those tasks that you can just start by putting terms into a spreadsheet, or into the SharePoint term store.
As taxonomy software developers, we see most classification schemes fall over after a year or so. Because it’s considerably more difficult to build classification schemes that sustain over time. Those schemes that persist do so because they meet meet the collective needs of the organisation.
So here are some of the questions you need to ask before you start. But finding the answers may be harder than you think.
Who are the stakeholders for the classification scheme?
You can’t just assume that your frontline users (those who are classifying information in your system) are the only stakeholders. Many others will have a vested interest in your scheme:
- Searchers both within and from outside your information ecosystem who want to find information
- Data analysts who want to mash your information with other databases
- Records managers who have to appraise the information for retention and disposal – if not now, then in a few years’ time Legal and compliance managers, with oversight of risk, access and security, privacy, and
- ICT who will have to migrate/archive your information into another system
You need to find them and consult with them.
Are there existing standards you should be adhering to?
All across the world, governments and industry are working toward data/metadata governance standards particularly wrt metadata formats, structure and values. Standards facilitate compliance with legislation, harmonisation across boundaries, interoperability and integration between systems, and support functionality such as enterprise search.
When you consult with other stakeholders, chances are you’ll find out what standards are applicable to your classification schemes. These include standards such as ISO 15489 Records Management and ISO 23081 Recordkeeping Metadata. Plus you should consult with peak industry bodies or your government regulators.
Are there industry classification schemes you can or should be using?
Depending on your industry or jurisdiction you may be required to use approved schemes with standardised terms. You may find more information on the websites of your peak industry bodies and government regulators.
We’ve seen contractors build lists of Commonwealth legislation and agencies from scratch because they didn’t know who to contact or where to source these schemes from.
Are you sure you’re not reinventing the wheel?
Everything that you’re about to do has been done before, in a different system, at a different time. Most organisations have established classification schemes – in network drives, in eCM systems, in SAP, in line of business systems. These schemes range in structure from simple lists to deep and complex file plans. Some are good, some are past their best. Many have been built at a considerable cost to the organisation by external consultants.
Maybe the scheme/s you are building already exists and just need updating!
What is the best information architecture to use?
Are you stuck in the hierarchical headspace?
Many folk still cling to the view of an all-encompassing file plan that forms the backbone of their eCM system. Originating with Windows Explorer, the hierarchical approach to classification transplanted itself into eCM systems as a system for arranging “records”. Given that the hierarchical approach to classification was the main game for over 20 years, it’s not surprising that many folk still view classification as a hierarchical event.
But times have changed folks. Contemporary information systems provide many more options for classification. For example, in SharePoint you can create a mix of classification options through team sites, sub-sites, libraries, lists, content types, folder structures, lookup sets, and managed metadata. Information architects can now tailor classification to better suit the needs of their users.
Of course such flexibility creates its own problems. But that’s another topic for another day.
What relationships will the classification scheme have with system processes?
Do you know which classification terms (and relationships) are used to signify or trigger system processes relating to:
- access and security
- retention and disposal
- legal holds
- data transfer
- other workflows
- migration into/integration with other systems
There’s an entire logical data model that connects classification concepts with information governance policies and workflow. It may not be well expressed but it exists, sometimes in the form of records retention schedules and security classifications. If you’ve consulted with other stakeholders you should be getting a clearer picture of information governance requirements.
How will your classification scheme to be maintained over time?
Even if you get else everything right, your scheme is going to change over time, starting from tomorrow. How are you going to manage the process of change, who is responsible, and are they across the answers to all of the questions we have just put to you?
Sustainable classification for information governance
Sustainable classification is a challenge but it’s not as dark an art as many suppose. When you understand the logic and follow the standards it becomes much easier over time. I’ll endeavor to explain further in my next few posts.