3 Key Takeaways from Enterprise Data World 2023
Brendan Healy
The Enterprise Data World conference is widely recognized as the most comprehensive educational conference on Data Management. At this year’s conference, the prevailing tone could be described as more cautiously optimistic than conferences of years past. To paraphrase speaker Donna Burbank from Global Data Strategy, Ltd., over the past decade, “the data people in a business have shifted from their company’s basement to the spotlight." Data has never been more of a buzz word, with 99% of Fortune 1000 companies reporting investments in data and AI, yet just 30% have developed a well-articulated data strategy. Thus, the messaging from EDW was clear, the need for an effective data governance plan is more important than ever.
According to the Data Management Body of Knowledge, data governance is defined as: “the exercise of authority, control, and shared decision-making (planning, monitoring, and enforcement) over the management of data assets.” In practice, data governance spans various fields across an enterprise, from business stakeholders and executive sponsors to data architects and data governance leads. All of these voices were prevalent at EDW this year, and while they would agree that there is no singular approach to data governance, there were a few key takeaways from the conference that pertain to any data-centric enterprise in the 21st century.
Takeaway 1 – Data Strategy Can Range from Offensive to Defensive
At a high-level, a data strategy is defined as a long-term plan outlining the people, technology, processes, and rules required to manage an organization's information assets. While this sounds straightforward, the business goals of an organization and the type of data it collects will dictate whether an offensive or defensive data strategy approach is appropriate.
On one hand, a shoe manufacturer might focus on an offensive data strategy that utilizes product data to improve profitability, increase customer satisfaction, and gain a competitive edge. However, a health insurance company is more likely to implement a defensive data strategy in order to comply with regulations, keep enrollees’ data secure, and mitigate risk.
As was the case with many of the companies present at EDW, an enterprise will likely have various business goals which land it somewhere in the middle of this data strategy spectrum.
Takeaway 2 – Be Thoughtful in Assigning Stewardship of Data
Within any business, the Data Steward is the individual or individuals responsible for the day-to-day management and quality of data. The assignment of data stewardship is unique to every organization and how their business is structured. Whatever the approach may be, there should be no ambiguity about who is serving as the data steward.
For instance, an e-commerce company may take a system-centric approach to data stewardship, where the steward is responsible for all the data within a certain system (CRM, CMS, ERP, etc.). On the other hand, a non-profit may take a capability-centric approach to data stewardship, where stewards are accountable for key data within certain domains of the organization, such as fundraising, marketing or administration.
Regardless of the situation, it is important to avoid taking an ambiguous approach to data stewardship, where it could be unclear if sales data is owned by the salesperson, the finance team, the billing system owner, or others. No matter how the stewardship model is broken down within an organization, there should be a clear understanding of who is maintaining the quality and consistency of that data for when it comes time to utilize it.
Takeaway 3 – Do Not Underestimate the Importance of Data Quality
It is estimated that data scientists spend a majority of their time cleaning and preparing data when they could be deriving insights from it. This issue can be better addressed across the data lifecycle, starting with the foundational elements of your data strategy. A concrete metadata schema establishes clear business definitions for all data being captured while an entity-relationship data model visualizes how the data is related. This entails answering questions like: Is a prospect someone who filled out an interest form, or simply opened the email? Does the customer_email field fall under the customer_name field, or vice versa?
Once data is being captured, a data quality dashboard can be utilized to measure key performance indicators (KPIs) like data completeness, consistency, timeliness, and accuracy. While it may be difficult to see the immediate return on investment from these steps, they go a long way towards ensuring a business has high-quality data to derive insights from.
Conclusion
In summary, the primary takeaway from the conference was that mindful data management applies to everyone, as demonstrated by the wide range of industries and organizations present at EDW. While there is no one-size-fits-all data strategy or data governance plan, there is no doubt about their necessity in our data-driven world.