Digital commerce has changed the face of retailing forever. A combination of wide product choice, speedy delivery, and simple-to-use search functions, that also recommend related products, have raised the bar.
While big data lies at the heart of digital commerce success, organisations often have trouble extracting similar benefits from their own information. They shouldn’t as a data governance program make it as easy to find the right data to manage their business as it is to shop online.
Getting data governance right brings rewards but equally, getting it wrong entails real risks. Poor data quality costs organisations an average of $US15 million a year in 2017, according to Gartner research. Meanwhile, almost half of CEOs (45%) say their customer insights are hindered by a lack of quality data, according to a KPMG report.
To meet the challenge requires a deep understanding of an individual organisation’s data landscape, which can be revealed by taking a logical and structured approach.
1. What data do I need?
Not all data is created equal. Data consumers have a huge role to play in an organisation’s data journey by precisely describing the type of data they need to perform their roles.
At JPMorgan for example, internal data consumers need to define their data carefully using enterprise-wide terms to ensure that data providers understand what data is required.
2. What data do I have?
Each organisation must audit the data it holds and ensure it is described consistently. This may involve creating a business-friendly language to describe your data landscape.
Accurate data descriptions can help people such as data scientists, risk managers and modelers understand what data exists internally and where to find it so they don’t end up needlessly asking other staff or bringing in more data sources.
At JPMorgan, we have created a tool called DNA (Data Network Architecture) where we apply descriptive tags to our data which can be used for queries to locate and interrogate such data.
3. Data lineage: where does my data come from?
Many organisations consume so much data they lose track of where it comes from and goes to. An audit may reveal that large and complex organisations are sourcing the same data multiple times or similar data from multiple vendors.
Understanding data lineage can help technologists and business owners decommission redundant or outdated systems. Regulators also want assurance that organisations can trace their data back to its points of origin to ensure accuracy.
Knowing the true source of all data can speed up time to market and is also the first step toward improving data quality. In cloud-based systems where storage is paid for based on usage, avoiding duplication leads to cost savings as well as greater efficiency.
4. Where should my data come from?
Knowing the source of data prompts the question – is it coming from the right place? Using specifically ratified authoritative data sources helps ensure better data quality and decision making across the business.
It is worth considering roles such as data content owners who are responsible for discrete sets of data, and for providing a data control summary that describes controls that support data quality, security and resilience. This is crucial because while a successful data governance strategy eliminates many risks, it also introduces a potential single point of failure.
5. How much reference data are we sharing?
Reference data, such as customer data, product data, instrument data, or pricing data, is the lifeblood of many organisations.
Centralising reference data can help make organisations more efficient. For example, if client data and documentation are managed in one source, an organisation can understand everything they hold about that client and access that across functions to provide seamless, informed service.
Data governance is a continuous journey and focusing on real-world business problems with a structured approach can achieve results.
Rob Casper, JPMorgan chief data officer
