Data management best practices: a practical checklist built on DAMA's framework
Data management covers eleven areas in the industry's standard framework. A mid-size company doesn't need all of them at once. Here is where to start.
The most widely used reference for data management is the DAMA-DMBOK, the Data Management Body of Knowledge published by DAMA International. It splits the field into eleven knowledge areas:
| Knowledge area | What it covers |
|---|---|
| Data governance | Who decides, who owns which data, and the rules |
| Data architecture | How data flows between systems |
| Data modeling and design | How data is structured |
| Data storage and operations | Running databases and storage |
| Data security | Access, privacy and protection |
| Data integration and interoperability | Moving and combining data between systems |
| Document and content management | Files and unstructured content |
| Reference and master data | Shared core records such as customers and products |
| Data warehousing and business intelligence | Reporting and analytics |
| Metadata | Data about data: definitions, lineage, owners |
| Data quality | Accuracy, completeness and timeliness |
A mid-size IT team can't build a program for all eleven at once, and doesn't need to. The practices below are grouped by what to do first.
Start here: ownership and definitions
- Name an owner for each core dataset. Customers, products, revenue and employees each need one business owner who decides what the data means and who can use it. Without owners, every other practice stalls.
- Write down definitions for the numbers people argue about. What counts as an "active customer"? When is revenue recognized? A one-page glossary ends more disputes than any tool.
- Keep a simple data inventory. List the main systems, what data each holds, who owns it and whether it contains personal data.
Then: security and access
- Grant access by role, not by person, and review it every quarter.
- Classify data into a few levels, such as public, internal, confidential and regulated, and apply controls by level.
- Know where personal data lives. Privacy laws such as GDPR and CCPA assume you can find and delete a person's data on request.
- Back up critical data and test restores. A backup you have never restored is a hope, not a control.
Then: quality
- Measure quality on the data that matters most. Track a few checks on core datasets: missing values, duplicates and stale records.
- Fix problems at the source. Cleaning data in reports hides the cause. Fix the form, integration or process that created the bad record.
- Set up one master record for customers and products before adding more systems that each keep their own copy.
Then: integration and reporting
- Document data flows between systems, so you know what breaks when one changes.
- Prefer standard integrations and managed connectors over one-off scripts nobody maintains.
- Publish certified reports for key numbers, so leaders aren't comparing different versions of the truth.
Keep it going
- Record lineage for important reports: where each figure comes from and what transforms it.
- Review the program twice a year. Retire data you no longer need, check owners are still in role, and update the inventory.
DAMA is also working on a third edition of the DMBOK. The knowledge areas above come from the current framework.
What IT teams should do
- Name business owners for your four or five core datasets this quarter.
- Draft a one-page glossary of disputed metrics and get the owners to sign off.
- Run an access review on the systems that hold confidential and personal data.
- Pick three data quality checks for your customer data and start tracking them.