Demystified’s Data Governance Principles
In digital analytics, “Governance” is a term that is used casually to mean many different things. In our experience at Analytics Demystified, every organization inherently recognizes that governance is an important component of their data strategy, yet every company has a different interpretation of what it means to govern their data. In an effort to dispel the misconceptions surrounding what it means to truly steward digital data, Analytics Demystified has developed seven data governance principles that all organizations collecting and using digital data should adhere to. These principles constitute a thorough consideration of stewardship of digital data throughout its lifecycle. Organizations that adopt and apply Analytics Demystified’s Data Governance Principles can operate with the assurance that they have a solid program in place for managing digital data.
The following principles constitute a responsible data governance program:
1. Collection
– All organizations collecting data across digital platforms must be aware of exactly what data they are collecting and how they are attaining that information either directly, through user agents, or via third parties. Data collection methods should be cataloged and documented to identify any data that is extrapolated, passively collected, or explicitly collected on web pages, mobile sites, apps, and other owned digital media assets. Further, this documentation needs to include information specific to the technologies employed such as log file processors, web analytics tools, panel based trackers, tag management systems, and other solutions used to collect all types of digital data.
2. Quality
– Data quality is critically important when governing data for business use. The first component of ensuring data quality is to audit data collection agents to ensure that data collected is in fact what an organization believes that they are collecting. In our experience, we’ve recognized that most web analytics implementations devolve over time. This often leaves organizations with data elements that do not align with business requirements, do not function as designed, or those that have been obfuscated by technology without any clear indication of what the data represent. We advise companies verify data collection implementations and to regularly audit their data to ensure data collection tags (if used) are firing properly and that existing tags are not producing duplicative data. Further, we advocate for routine data quality checks to validate ongoing data collection and to alert organizations to potential data collection errors.
3. Access
– As companies implement data collection methods and provision access to employees and potentially contractors, agencies, and technology partners, access to an organization’s data becomes an increasing concern. The first line of defense for governing data access is to only provision access to email accounts of corporate employees or trusted agency partners (i.e., no personal or gmail accounts). This is an easily administered best practice that reduces the risk of a former employee gaining access to your businesses’ data. A more challenging aspect of data access that you need to govern is when technology partners share data with others. Often this is aggregated, non-identifiable data, yet organizations must be aware of instances of data sharing by third parties, data aggregators, ad servers, targeting technologies and other solutions that potentially compromise restricted access to your digital data.
4. Security
– Data security is often coupled with data access, but we at Analytics Demystified believe that data security goes beyond merely provisioning access to qualified analysts, but that it includes a business’ ability to safeguard its data stores. Most data collection solutions today are amassing large volumes of data and have opted for cloud-based storage solutions. While nearly all of these solutions fortify their security with multiple layers of redundant measures, the onus of understanding where and how any data is transferred from these solutions to other technologies falls upon the business. An area of concern is data “leakage” that could occur when data is inadvertently (or unknowingly) shared with external parties. Companies should minimize this risk by clearly understanding and documenting how their data is stored, shared, and secured across all data collection agents.
5. Privacy
– For any business that is collecting consumer data, privacy is a critical concern. It is the responsibility of the business to inform consumers what data is being collected and how that data will be used. Numerous best practices exist around divulging this information within published privacy policies, but the best guidance that we can offer is to deliver a clear and concise data usage/privacy policy that offers an opt-out for consumers who do not wish to be tracked. Businesses should also be aware of and classify data that is anonymous, segment identifiable, and personally identifiable and treat each independently. This classification element of data governance should be governed at the technology level because it extends beyond web analytics technology into business intelligence, enterprise marketing management, and customer relationship management solutions.
6. Integrity
– Governing data integrity is predicated on the fact that many data collection technologies today leverage processed data in their outputs. In web analytics tools, this might equate to a series of actions that constitute a “user session”, or a “path” the leads to a conversion event. In other technologies like IBM Tealeaf, multiple online activities can be associated with a single session that provides the necessary context for the data output. This data often requires that it is presented in processed form such that it reveals the true nature of what happened in a digital environment. Many businesses have the temptation to disaggregate processed data for inclusion in an enterprise data warehouse as “raw” data that can be analyzed at a later date. However, there are inherent risks in doing this because it could lead to inflated activity counts, incongruous data, or simply incomprehensible data. For these reasons, data integrity is an important principle of governance to ensure that data is utilized and analyzed in its intended form.
7. Presentation
– In digital analytics, there is an old adage that by “torturing” your data you can make it say anything you want. Responsible stewards of digital data are cognizant of this fact and strive to present data in proper context. While some organizations attempt to assign gatekeepers to assure data is presented in proper context, this becomes increasingly difficult as data sets accrue to petabytes in scale and access is granted to numerous individuals. Rather than restrict access to a responsible few, Analytics Demystified recommends companies to train their employees to recognize what data is being collected and what outputs are appropriate for specific data types. This level of education minimizes improper data interpretation and is the foundation for solid presentation and delivery of digital data assets.
It’s important to note that these Data Governance Principles are merely a starting point for developing your own data governance program. In our extensive experience consulting with organizations of all sizes, each presents their own data governance challenges. By adopting these seven principles and institutionalizing a process around each, companies can operate in today’s increasingly digital world with the confidence that they are responsible stewards of digital data and that they have taken precaution to safeguard their data according to industry best practices.
To learn more about any of Analytics Demystified’s Data Governance Principles, please reach out to us at Partners@analyticsdemystified.com, we’d love to help launch your data governance program or to learn how you’re currently governing your digital data.
