Believe it or not, there really is such a thing as too much data.
Or rather…too much poorly-managed data.
Recently I was on a project for a major music label that was struggling with a variety of data-related issues that were materially impacting the business. Their concerns were that data was not maintained correctly, there were delays in artist payments and in product releases, and they’d had to deal with lawsuits from incorrect royalty payments, among other issues.
My team was brought on with the goal to improve the quality of the data they used to run the business and to create a practical data governance strategy.
Understanding the Bottleneck
The first step to implementing any data governance process and related best practices is to fully understand the current people, processes, and tools used in data maintenance. You can’t fix what you don’t understand, so getting a solid picture of the data lifecycle up front is key. The music industry, like most media and entertainment industries, is an event-driven industry. Most of the new release setup process happens sequentially, but there are many dependencies and opportunities for different attributes to be maintained in parallel.
Communicate the need and the value proposition of this taxonomy exercise up front, and get the right support across the organization. We did just that, and secured stakeholder and executive support from both IT and business teams, which put us off to a great start.
Once we understood where the bottlenecks cropped up, we had to make sure everyone saw the same picture. Before they could apply better data governance strategies to reduce these issues, our key stakeholders—the label heads and the supply chain executives who lived with the daily pain points—needed to understand where in the flow these problems were occurring. Then, in order to analyze those potential failure points, they had to better understand the end-to-end flow of their key business data, beyond their immediate groups, and how it was used by the entire business.
To get there, we first documented all the key organizations and departments, focusing on the key players involved in the maintenance and consumption of data. Take note: unless this has been done previously at your organization, this exercise can be lengthy and painstaking, as it requires quite a bit of detective work and numerous interviews and triangulation of information.
After the taxonomy and ownership were firmly determined, we were finally able to build a data model of key data entities that were mapped to key business users and use cases of each data entity.
What Does a Day in the Life Look Like?
Now that we had a map, the next step was to survey the territory. We used a “day in the life“ approach to define where each key data entity and element intersected with different business processes, then pieced together those business processes and sequences into end-to-end business scenarios.
(A practitioner note: taking a data-centric view to get to business scenarios is merely one vantage point; other lenses such as process efficiencies, role matrix, and material or legal flows can be used as well to create end-to-end business scenarios).
During this process, we also captured key pain points for each group and the related business impact of those issues on the overall value chain.
For the recorded music, the major events would look like this:
Step 1: A&R (artists and repertoire) finds an artist
Step 2: Contracts are developed
Step 3: Content is created
Step 4: Products are created from the content
Step 5: Products are sold or licensed
Step 6: Payments are received from customers
Step 7: Royalties are paid to the various participants
Each of these events was then decomposed into a set of sub-events and, in turn, for each sub-event, we further documented the people and systems involved and the key data entities/elements impacted.
After we collected all these details, we generated an intricate network of interdependencies that wove together this new data governance model with the business process and organization, systems, and data use cases across all of the groups that had touchpoints with the data entities.
Seeing the Bigger Picture
What we found in looking at the entire puzzle was that, while different groups might have well-defined processes, the overall data lifecycle process across all of these groups was not very well understood and, as a result, not effectively controlled. Groups tended to focus on the data entities/attributes most critical to their specific business area (naturally) but in addition, were frequently and, unbeknownst to them, further responsible for entering other data elements that were critical to other business areas further downstream.
There were holes in clearly defining who owned what data and this lack of clarity created issues because the wrong people ended up maintaining and making decisions around the data during the lifecycle. Error, churn, rework, and frustration were the unintended consequences of this disjointed way of working. We knew that to fix this problem, it would be important moving forward to be extremely specific about assigning data to certain people who would then be responsible for it.
Developing the Data Governance Framework
Now that we had the view of the people, processes, and tools being used to maintain the key data entities across the business, as well as a compiled list of data issues, we were able to define a practical data governance framework to support the organization moving forward.
First, we identified and established data stewards within each department to address the issues around the lack of ownership and accountability for the end-to-end data maintenance. Next, we developed a scorecard process to continuously monitor the health of each of the key data entities.
Second, we coached the newly established data stewards on how to use the scorecard and evaluation process to take a snapshot of the organizational use, processes, tools, and ultimately the overall data quality for each key entity, which provided the organization a baseline of where they were and a way to measure improvements going forward.
Finally, we deployed an Agile data sustainability process where the teams could collectively across the various departments maintain and prioritize various efforts required to implement the framework.
A Long-Term Practical Data Governance Strategy…Success!
Data governance is not a project; it is a discipline. By developing a systematic approach to managing the data lifecycle and by maturing the organization to operate from a single global backlog of prioritized key improvement efforts, the business was able to budget and assign resources to focus on the key data entities and issues that provided the greatest business value.
With this approach, our client, and truly any organization grappling with data challenges, is able to quickly address key data issues impacting the business while continuing to drive long-term data governance maturity across the organization.