An organization looking to standardize its data should consider the following four steps as a guideline for their process. As organizational needs vary, the specific implementation of these stages may change, but the basic procedure should remain the same.
Step 1: Conducting a data source audit
In the first stage, organizations should mark all of the data sources (supply points) feeding them information. This step requires many organizations to break up informational silos that may exist between departments; moving forward it will be key to loop in all decision-makers for a project to ensure they are getting a full picture.
In general, the more data, the higher level of accuracy an organization can depend on. However, there are important caveats about the compatibility of this data. In order to build good data standards, an organization must know all of the kinds of data flowing in, how often it is updated, which teams use specific data logs, and the nature of the data source. The answers to these questions will prove critical for success in the next stage of data standardization.
Step 2: Brainstorming data standards
This step involves a level of creativity on the part of the organization’s decision-makers. A positive proportional relationship exists between the amount of data collected and how precise standards must be; in other words, the more data an organization has, the more important it is for effective standards to format that data.
Big data functions by the Three V’s: volume, velocity, and variety. By definition, this refers to a large amount of data arriving at frequent intervals that exists in multiple types. To sort this data, standards must be both all-encompassing and specific enough to hold up when the velocity or volume of that data changes. For that reason, organizations should be forward-thinking and account for the development of their enterprise. Standards must both organize the data that exists now and might exist in the future. For inspiration, some existing data standards exist. For NGOs, using pre-set standards might be advantageous for comparison with other groups in the same space, and the UN’s SDG indicators are a good place to start.
Step 3: Standardizing data sources
This step applies to the information constantly flowing into an organization. Before taking on the entire database, organizations should make sure all data supply points follow a common format in order to avoid replicating the same mistakes in the future. This step is crucial because significant differences in ranges used and measurement units can skew data analysis applications. These disparities make it difficult to receive reliable, predictive data because the inputs are artificially diverse. Organizations should keep the prevention of features with wide ranges from dominating their metrics.
One way to convert data to a common format is the usage of z-scores. This formula allows organizations to sort data points by number of standard deviations from the mean, and is calculated by subtracting the mean and dividing the standard deviation for each value.
Once data has been standardized in this way, all data points. will have a mean of zero, a standard deviation of one, and exist on the same scale. Another tool organizations can use to help merge formats is to use drop-down menus instead of free-form data entry; allowing free entry can create artificial variety by separating similar results (like creating separate categories for entries labeled “Road” and “Rd.”).
After creating a common format, organizations should add data validations to ensure that data is always collected in the same format and complies with the new standards.
Step 4: Standardize the database
The last step of data standardization involves implementing new standards on existing data. This stage requires a large investment of time and effort and can become quite repetitive. By retrofitting previous data entries to meet new standards, organizations ensure that each employee interprets data in the same way and operates on the same page as other departments.
Standardizing the database requires the implementation of filters designed to refine data sets by including specific information and excluding repetitive or irrelevant data. Creating good filters is key to standardization and exists in the same positive proportional relationship as standards and data; that is, the more data collected, the more important filters become. Good filters must be able to identify the details inside data points and make it possible to search and organize data sets by those specifications. For this reason, stage four is tied to stage two, as filters rely heavily on data input standards.
By investing the time and resources to standardize their data, social service organizations can significantly reduce costly overhead, expedite project management, and combine inputs from a variety of sources to improve their services. As data continues to redefine the nature of project management, standardized inputs and common database formatting can ensure your organization is equipped to keep pace with an evolving world.
If you’re interested in learning more about best practices in data standardization for organizations in your sector, please feel free to reach out to Anish Nagar (email: anishnagar [at] corecentra.com). Anish is the CEO of Corecentra Solutions, a software company providing purpose-built digital solutions for socially conscious and outcomes-focused companies, foundations, nonprofits, and frontline government agencies.
Corecentra provides advanced digital tools for organizations to manage, monitor, and report their social performance and impact. We help socially-conscious companies, impact investors, foundations, nonprofits, and frontline government agencies manage portfolios and programs, aggregate and analyze data, and easily report outcomes to key stakeholders. By seamlessly integrating program management, budgeting & finance, stakeholder engagement, predictive analytics, and impact assessment, our products empower organizations to increase their social impact and deliver a quantified view of social performance to investors, donors, beneficiaries, employees, and communities.
Our Emergency Response & Impact Management (ERIM) platform allows organizations to efficiently manage a wide variety of pandemic-related projects, including economic support programs, community health initiatives, medical supply efforts, and philanthropic activities. Instead of waiting for data to trickle in from legacy systems and processes, leaders can use ERIM to track results and make data-driven decisions to help communities in real-time.