LESSON - Building a Healthy Foundation for Strategic Business Intelligence
Strategic business intelligence (BI) initiatives and analytics depend on the quality of data used by the organization in its efforts to improve performance, discover new opportunities, and operate more efficiently. Data quality issues can be a fatal flaw in BI implementation.
By Bud Walker, Director of Data Quality Solutions, Melissa Data
Strategic business intelligence (BI) initiatives and analytics depend on the quality of data used by the organization in its efforts to improve performance, discover new opportunities, and operate more efficiently. Data quality issues can be a fatal flaw in BI implementation.
Industry analyst firm Gartner, Inc. reports that more than 50 percent of BI and data warehouse projects suffer limited acceptance and are more prone to fail due to poor data quality. Simply put, poor data quality results in poor quality decision making—so the most critical part of BI implementation is to capture and maintain timely, relevant, and trusted data.
Implementing Data Quality: The Prescription for Robust BI
Evaluating and maintaining data quality is complex, especially in the healthcare sector. Data streams in from multiple touch points and channels. Data entry errors can occur during patient registration, as patients move, when physicians change practices, and so on. The key is to develop a process that prevents bad data from entering your database in the first place, and then enrich accurate data for a deeper understanding.
Verify Your Data
Contact data is one of the most important assets of your enterprise, and necessary for effective BI initiatives. Develop a system that corrects and updates incoming data in real time to verify and standardize street addresses, phone numbers, e-mail addresses, and full names. Having clean, consistent, and standardized contact data will facilitate data aggregation, analysis, data mining, and duplicate detection.
Match and Dedupe Your Data
Identifying and preventing duplicate data is important for patient care, and a vital step in BI implementation. Many systems can only deal with exact matches, so something as minor as a misspelled name or a record entered as “Beth Smith” instead of “Smith, Elizabeth” could yield separate medical records. When medical staff members try to search these records, previous medical histories, x-rays, and laboratory and test results may or may not be available, depending on how the data was entered.
The key is to develop a process that prevents bad data from entering your database in the first place, and then enrich accurate data for a deeper understanding.
There are robust deduplication tools available that employ fuzzy matching algorithms to identify these difficult-to-spot, non-matching duplicate records. Eliminating duplicate records and merging or purging them for a unified view of the record will not only result in cost savings, but also improve clinical care and decision making.
Enrich Your Data
Enriching your data can fill in gaps for complete contact information, including missing e-mail addresses and phone numbers, or updating records with current address information. Enrichment can also add deeper meaning to your data by appending additional information to your records such as street data (residential/business indicator), phone numbers (time zone), e-mail addresses, geographic information, and demographic data (NAICS code, household data, property information).
Geocoding addresses will append the latitude and longitude coordinates to gain location intelligence, including census tract and block numbers, county name and FIPS, and Census CBSA statistics. This enhanced information improves the accuracy of any process or business decision that depends on location data, including market segmentation, mapping, risk assessment, logistics, and service eligibility.
Gender coding or genderizing your data provides greater insight into your contact data by identifying the record as male or female based on the first name. Genderizing allows you to filter and analyze your data based on gender, often an important demographic, and that information can strengthen your BI efforts.
The End Result
Integrating a continuous data quality process that blocks bad data from entering your system and enriching clean data for deeper insight is the foundation for building a robust BI platform. The end result is better decision making, increased cost savings, and improved operational procedures.
This article originally appeared in the issue of .