As data stores grow in size, successful data management becomes an issue. In the past, data was perceived as the by-product of a business activity and it had little use once the process was completed.
Today the results of reporting and analytics have made the use of data desirable for many new business initiatives. It is common for application data to be shared with a number of other systems. Organizations, therefore, need to build comprehensive data management strategies that address today’s realities.
What is a data management strategy?
Data has become a critical asset to support informed decision-making and a data management strategy has to improve ways to acquire, store, manage, share and use data across an enterprise in a way that’s repeatable and efficient. A data management strategy is, therefore, basically like a roadmap for an organization to use to achieve its goals.
Companies that use an operational data store can aggregate data from multiple systems of records and obtain a more complete view of data that enables more comprehensive reporting. This is one of the ways they can support informed decision-making and avoid common data challenges, like a siloed view of data.
Common data challenges
With an effective data management strategy in place, it is possible for organizations to avoid common data challenges. One of these challenges is that most organizations use different systems of record and reporting on each data source separately offers a siloed view of data.
Most application systems were built as individual data processing engines. They contained the data needed to perform defined duties and there was hardly any thought given to sharing data across applications. The organization and storage of data was for the convenience of the specific application that collected, created and stored the content. This is not practical today where dozens of systems rely on data from multiple sources to support individual business processes.
Other problems organizations can avoid with a proper data management strategy are incompatible, duplicated or missing data and data activities that use time and resources but don’t contribute to the goals of the business.
Identify business objectives
It makes no sense to collect, store and analyze the wrong types of data. This is why business objectives need to inform a data management strategy. It helps organizations to focus on some critical use cases for data in the business and build a strategy accordingly.
Collection of data
Knowing what to use data for helps to determine how to collect, prepare, store and distribute it. When thinking about the collection, it’s necessary to decide on factors such as which data sources to use, whether to use unstructured, structured data or a combination of both and how to collect data.
Preparation of data
Establishing guidelines for naming data, identifying incomplete or disparate data and deciding how to clean raw data is all part of the preparation process.
Libraries use card catalogs as it is impractical to try and remember the location of each book. Metadata is important for business data usage because it is impossible to know the location and meaning of all the thousands of data elements across many data sources. A data strategy must ensure that all data assets can be identified.
Processing of data
Data generated from applications is a raw commodity and certain activities are necessary to transform, correct and format it. In most companies, data comes from both internal and external sources. Internal data is generated from many application systems and external data may come from cloud applications, business partners or government agencies etc.
Making data ready to use is about offering tools and establishing processes that individuals can use without having to rely on IT involvement.
Advanced technology is easing business processes globally in so many ways today.
Storage of data
Businesses are dependent on sharing and distributing data to support operational and analytical needs. When each system creates its own copies of data, this considerably increases storage and processing needs. Therefore, it is critical to store data in a way that simplifies access without requiring everyone to create their own copies.
An operational data store (ODS) centralizes data from various sources into a single location where it’s available for business reporting. It provides the most recent relevant snapshot of data from all the transactional systems. Previously isolated systems feed into the central database offering a consolidated view of data. It is much easier for users to diagnose problems without having to go to the component systems. An ODS also has time-sensitive business rules that greatly improve business efficiency.
An ODS has a different purpose than a data warehouse. It performs simple queries on small data sets, whereas a data warehouse performs complex queries on large datasets. Businesses use the centralized repository of a data warehouse to inform strategies for the whole business.
Data governance
The purpose of data governance isn’t to limit access to data or interfere with usage. Whether it is determining security, quality, privacy, data naming standards or establishing new data rules, effective governance ensures consistent management and manipulation of data and access to it.
Policies and procedures must be understood by everyone in the business and all employees need to know how to successfully execute their roles. Data governance offers the necessary control over data as changes occur to the technology, processing and methodology areas associated with it.
The bottom line
The increase in source applications and cloud-based applications means there are just so many systems, sources and data to track and manage. Moving forward with a data management strategy involves identifying strengths and weaknesses that exist within a business and setting achievable and measurable goals to work towards improving data access and sharing.
This is not a one-time initiative but an ongoing one where it is necessary to review and measure progress on an ongoing basis. Businesses will not be successful in their efforts to improve the management of data unless they address every individual data-related component.
