Declutter Your Data: tidy up to maximize value (3/3)
After my previous post Get Ready to Tidy, I hope you’re excited and optimistic to get your data home in order. Now we start the tough, but highly rewarding process of evaluating, deciding, and taking action.
Decide What to Do
Reinforcing the power of simplicity, Kondo presents a single decision point for tidying up. Once you determine if something sparks joy, you either keep it or discard with gratitude. This is not just ritualistic fluff though. It’s a moment of reflection. You learn something about yourself when you express gratitude. You say Thank You to the item as you place it in the box to leave your home.
In the same way it’s important to reflect on our data assets regularly as well. Why did I collect this data in the first place? What value has this data generated over time? Does this data still add value?
Kondo’s full process for tidying up includes:
Fetch every item in a specific category from every part of your home.
Bring back every item to a single location and group them together.
One by one, touch every item and consider: Does this item spark joy?
If an item sparks joy, set it aside to keep.
If an item does not spark joy, put it in a box to leave your home.
Here’s my equivalent process for decluttering data:
Identify every data asset in a certain category from every part of your organization.
Document all data assets in a single location, such as a spreadsheet.
One by one, focus on each data asset and consider: Does this data asset create value?
If it creates value, mark as Keep.
If it does not create value, mark as Retire.
Organize & Catalog
Once you decide what to discard, Kondo has tips and tricks for organizing everything that remains, such as:
Designate a specific spot for each and every thing.
Store all items of the same type in the same place.
Make sure everything is visible and accessible.
Unpack and de-tag new clothes immediately.
Here’s how I translate these tips for data asset management:
Make sure every data asset is clearly owned and managed by a single group or team in your organization.
Catalog every asset in a single system by category: Data Pipelines, Data Sets, Data Tools.
Make sure data assets are easy to find and use, especially by searching.
Classify and catalog new data assets immediately.
Thoughtful classification is essential for organizing data assets. At a minimum, classify and document the following info for every data asset you decide to keep or create in the future:
Name | Short descriptor of what it is.
Creator | The person who originally created it.
Owner | The person who owns and manages it now, whether they created it or not.
Group | The group at your company that is ultimately responsible for its health.
Purpose | Its reason for existing. How does it create value for your company?
Documentation | Links to design docs, diagrams, code, or any other information explaining how it was built and appropriate use cases.
If you decide to retire a data asset, I recommend the following process:
Turn off automated refreshes.
Set an expiration date.
Move the data asset to an archive space in your data environment.
Delete it if it has not been used or created any value by the expiration date.
You may be thinking: “Data storage is cheap. I’ll just keep everything.” I’ll just remind you of this though:
It doesn’t cost much financially to keep all this stuff in this space. The mortgage or rent remains the same no matter how much stuff you put in it. However, it will cost you time (and money) finding the right data, wrong decisions using the wrong data, and overall satisfaction using your data.
You’ll generate greater impact from data that truly creates value if you let go of data that doesn’t.
