Five Steps to Improve Business Insight Through Data Quality
Organizations generate phenomenal amounts of data every day. To avoid being inundated and to leverage the full potential of what enters their business continuously, organizations must have a clear strategy for leveraging their data. New technology innovations accelerate the generation of data at an exponential rate; a few years ago no one was using social media, now every day billions of people post unimaginable amounts of data about themselves online non-stop – and that’s just people! Factor in data from devices (IoT) and the number quickly jumps to more than a trillion. Data at this level creates an unprecedented management challenge, forcing all organizations to adapt continually and rapidly in order to gain the insight needed to improve decision-making.
Businesses understandably demand more from their decision-support systems, which require high-quality data to function effectively. Leaders need this information to gain better insight into what is happening in their organization and in their industries, preferably as quickly as possible. Insight into the data deluge, in addition to up-to-date decision-support systems, are the key to success, but both rely on high-quality data.
Insight is defined as having a deep and accurate understanding of some phenomenon, trend, data or content. When leaders use decision-support systems, they are doing so to understand information dynamics to drive the organization in a positive direction. These systems, however, are only as good as the data they contain, which is why it is important to minimize excessive or irrelevant data. With excess comes confusion and contradiction, which can result in the overall quality of data being diminished and undermined. Excess data will make its way into the decision-support systems and ultimately may cause negative impacts on operations.
The following five steps should be taken to improve operational insight as well as to minimize the negative effects of bad data:
Reduce: Determine which systems, applications and technologies are generating operational data and determine where overlaps exist. Eliminate as many overlaps as possible.
Generate less: There is rarely a case where no data exists. It doesn’t mean, however, that what data does exist is good. Installing new tools and systems that create more data is not the solution if the data is not relevant, it only adds to the uncontrolled volume already in place.
Categorize it: One of the first steps to understanding is grouping or categorizing information into collections that individuals can search and analyze to harvest what they need. An environment with an overabundance of bad data is only made worse by allowing it to be an unorganized overabundance of bad data.
Priority and Precedence: At some point, you must decide which source of data is more valid than others. This does not necessarily mean it is completely accurate, but it should be more accurate than other potentially duplicate sources. Identify and document the sources and/or their contents with a relative prioritization scheme and share it with those who need to know.
Standardize & Normalize: After reducing the clutter and volume of data, the focus should move to normalizing and standardizing the data across the enterprise. This effort ultimately helps the organization to quantify the real value of improved data quality.
Each of the five steps above will help with improving data quality. The end goal is to improve operational effectiveness, and that comes with better insight into operational data. During improvement efforts, there is much to be learned about how the organization is currently operating by understanding how data and sources in use came into existence. Learning these lessons can help eliminate similar mistakes that could lead to high volumes of low-quality data. In the end, leadership must be able to make informed decisions based on a clear and precise view of what is working and not working within their organization. The only method to achieve this end result is to improve the quality of all the data in the enterprise that feeds decision-support systems.
The 451 Group’s report on The State of Enterprise Data Quality provides insights into how organizations view the need and application of data quality. Download the report here.