My post on the criticality of an information inventory (sometimes known as a data map) seems to have hit a nerve, so let’s continue in that vein.
The typical approach to preparing a data map or conducting an information inventory involves just a few simple-to-describe steps: identify key stakeholders; ask them what data they work with, where they get it from, and what they do with it; and present their consolidated answers (often in the form of a spreadsheet). However, as currently performed, this task too often is fundamentally flawed because it starts with what we think we know about the information we have, rather than what actually is.
Flaws in the Conventional Wisdom
We know our environment pretty well, so chances are we’ll do a decent job of choosing the people to talk to and the systems to investigate. But there are certain risks associated with this approach that can seriously degrade the value of our work:
- People can overlook systems they either don’t often use or actually forget about.
- Teams work in multiple silos and don’t always know which one they’re in at a given moment.
- Data migrates and gets renamed, ultimately becoming invisible when you try to find it.
- The process itself can be time and resource intensive as meetings are scheduled and rescheduled and rescheduled again, and staff struggles to add yet another task to their already heavy workloads.
All of which is to say that by the time you finish your inventory, it may be riddled with gaps or quietly wrong in ways nobody notices until it matters – as when you use the results to feed clean data into your AI, identify PII, migrate to the cloud, or update your retention schedule. The list of initiatives is long and important, and every item on it relies on a solid map to be successful.
The Better Way
There is a better way to go about this, namely to start by finding out what information there actually is, and where, and why, rather than asking humans to recall and report what’s in their heads and in their procedure manuals (which may not reflect reality, and my not even exist).
Happily, the likes of discovery tools and autoclassification solutions can surface information that no one remembered to mention (legacy databases, forgotten file shares, duplicate records hiding everywhere from email inboxes to the cloud, etc.). These can be incredibly effective and should be utilized to the extent your budget allows – though be aware that they too are often deployed based on people-supplied input that goes far beyond document types and locations to include basic elements like vocabulary and taxonomy.
Greater Completeness and Accuracy are Well Within Reach
The takeaway here is that while preparing a data map or conducting an information inventory involves only a few fundamental and straightforward steps, their typical reliance on human assumptions to start greatly complicates our ability to ensure completeness and accuracy. Much better to begin with an accounting of the information we can prove is under management and go from there – and better still, this isn’t as hard or expensive to do as it may sound.
—
Need to create or update a data map/information inventory? Ping me today and let’s make it happen!