Part One: Data Strategy Is More Important than Ever in an Age of AI.
This will be a multiple-part series on data strategy and how it is the precursor to data management and data governance.
Firms often skip many essential steps to creating a data strategy favoring data lineage/governance, usually for regulatory compliance rather than creating a holistic yet integrated vision for data. While practicality is always good, it can be at the bane of getting the most out of the firm’s data over time. When a data strategy does not guide data governance, this keeps data governance in a defensive position in general; it is a big mistake that keeps data governance in the basement of the organization, being perceived as a cost center and not the revenue and monetization driver that data governance can be.
Let’s start with what a data strategy is and why your organization needs one. Then we will discuss in future articles how data governance needs a tighter connection to strategy.
Data strategy and how I like to think about it is a sharp vision for how your data is organized and turned into knowledge throughout the organization.
There is data, information, and knowledge. Each of these has some organization of data and planned use cases. I like this pyramid or hierarchy paradigm for data strategy. As you go higher, it’s about generating insights and improving the quality of decisions based on clean fit-for-purpose quality data.
20 Key Considerations In Your Data Strategy
Some key considerations in your data strategy, and I will not prescribe the answers to these considerations here:
1) How do you define quality data
2) Who gets to move data and to where?
3) Is there a planned level of data duplication, or is it, as they say, the “Wild West” with replication all over the place?
4) Do we want to have one version of the truth or multiple versions of the truth? What are the risks and benefits of each?
5) Are we using an ETL process or ELT in the age of big data
6) What types of data models are we using? Logical layers (star Schemas) no SQL, blob storage?
7) Are we using open gardens or data lakes, or a pond approach?
8) How do we define our data fabric at the firm?
9) What newer tools do we use for moving data. Are we using AI-based tools (RPA, etc.)?
10) Who can access PII or NPII data, and how do we create highly secured data zones?
11) How many self-service analytics tools do we allow? Do we need both PowerBI and Tableau?
12) Do we have an on-prem cloud approach or a full-on cloud data strategy?
13) Where do AI and cognitive technologies get their data
14) Do we have transparency in business rules and algorithms that drive our business?
15) How do we monetize our data, and at what point in the data lifecycle?
16) How many customer keys or unique identifiers do we carry?.
17) What is the role of generative AI?
18) How do we resolve the identities of both prospects and customers?
19) Who owns the data Do we have producers and consumer-defined roles?
20) Do we have a centralized or decentralized approach to data management, and is our organization clear about how we operate?
There is no formal data strategy if the firm doesn’t have clear answers to many of these questions. In addition, a data strategy is not a data management framework, which would come next once you have defined the strategy.
“Why is this important?” you ask. It will help you set the priorities for data governance and data management organizations (DMOs), rather than just having them fall into a project or two and perhaps only viewing them as the people who handle compliance issues or controls. Remember DMOs, enable data science, marketing automation, AI, CRM, and many other revenue-generating functions. An integrated enterprise data strategy will allow you to scale your data management and governance efforts, making the work more important and meaningful and increasing the focus on the business objectives and ROI.
I look forward to your thoughts on why you think the tail is often wagging the dog regarding data governance versus data strategy.
In our next issue, I will discuss master data management and data governance in detail.
