The 5 Ws of Data Management
As with celebrity divas or doing your own taxes, the challenging reputation of data management precedes itself. At an enterprise level, tackling a problem that involves this task can be downright intimidating.
First and foremost, anyone working for an enterprise company can attest that the term â€śdata managementâ€ť means different things to different people â€” even to coworkers within the same organization. Broadly, data management involves all the disciplines required to manage data life-cycle needs and to treat data as a valuable resource.
When applied to enterprise level projects, this broad definition certainly leaves room for interpretation regarding how a data management project should be approached and how to measure its success.
While data management might be complicated, you donâ€™t have to overcomplicate your projectâ€™s initial approach. Starting with a general exploratory framework is the first step to defining the projectâ€™s goals and metrics of success. Perhaps even more important, an initial structured list of general questions will prevent you from making assumptions about the challenges you might encounter along the way.
Thatâ€™s why I like to take a â€śWho, what, where, when, whyâ€ť approach to data management projects:
Who is in charge of the data?
Unless this question is answered with a clearly documented management plan outlining ownership polices (Hey! It could happen!), a lot of time will be spent understanding general ownership and the chain of custody. In some cases, data management is wholly assigned to the IT department because of its technical nature. However, in a different scenario, the data management process could be a cross-departmental quagmire.
Donâ€™t be afraid if your response to the above question is something like, â€śI think X owns Y; Iâ€™m pretty sure A might help with B; and C might own anything that affects Dâ€™s department.â€ť Take this as an opportunity to slow down and really examine who owns what.
What companywide KPIs should the data support?
Defining which key performance indicators the data should support is the first step to testing a data management process. Good-data quality will mean different things to different departments, and naturally, most departments should focus on different KPIs in order to service different goals.
However, interpreting the same data with different motivations can lead to disparate interpretations. Because of this, you should define both companywide baseline KPIs and individual ones that help departments achieve their unique goals.
Where is the data managed, shared, and reviewed?
Documentation of the existing data architecture is a great asset, but it shouldnâ€™t be relied on without further examination. When the data management process spans multiple organizational teams, planning might occur with one team, and collection, validation, and analysis might be the responsibility of several others. Even if the process is managed by one team, itâ€™s likely that the team will use an array of third-party tools, spreadsheets, and internal processes to manage the data life cycle.
Defining where data is stored and managed from â€śrequirements to retirementâ€ť will often uncover areas where the process should be refined or streamlined. More important, it ensures everyone is on the same page when changes are implemented.
When does data management need to be adjusted, and how fast are changes implemented?
The bigger the team and the more rigid the business processes, the slower the response is to new requests. When a company struggles to quickly tackle requests, its ability to implement improvements and test new methods is also limited. Determining the timing needed for changes will help frame your projectâ€™s approach and set the right expectations.
Why is this data valuable?
Modern companies should treat their data just like any other tangible asset. The value extracted from the data, however, will be different for different teams. Business teams will use data to make strategic decisions, customer support will use the information to gauge product satisfaction, legal teams will need to access data for regulations or litigation support, and finance departments might need to access data for audit purposes.
Asking â€śwhyâ€ť each department values data will ensure the right tools are used and the total cost of ownership is shared among the teams throughout the process.
Data management is by no means an easy task, but it certainly doesnâ€™t have to be as daunting as we make it out to be. Asking yourself these five questions will help provide the right perspective, narrow your scope, and streamline your initial efforts.
Once the answers become clear, youâ€™ll realize that data management isnâ€™t something you should dread as much as tax season; itâ€™s something that makes your business stronger and more successful in the long run.