In December 1998, NASA’s Mars Climate Orbiter launched toward Mars. It was slated to become the first satellite to observe weather and atmospheric conditions on the red planet. But before it made any observations, the satellite lost communication with NASA scientists. It soon became apparent that upon entering the Martian atmosphere, the Mars Climate Orbiter disintegrated.
Why? Bad data. One piece of software made its calculations in metric units, while another used imperial units. Newtons vs. pounds. And because of that error, a $125 million satellite vanished from the night sky.
You may not be working with projects the size of NASA’s, but your organization is also walking the fine line between opportunity and risk. And just like NASA, your business decisions are only as good as your data. Data holds enormous potential, and it’s easy to assume that the more data you have, the better. But the truth is that data is not all equally accurate or reliable. In fact, data quality studies conducted by GS1 US, a globally recognized data standards organization, found that up to 50 percent of data from the companies they surveyed was inaccurate.
If the data is inaccurate, then the decisions based on that data will be as well.
Most companies have more data than they know what to do with, but if they can’t access it or can’t ensure its integrity, then they will never benefit from the powerful potential it holds.
So where do you start? How do you take an underperforming system and turn it into a data powerhouse?
You have to start with quality data.
How Data Quality Drives Business Performance
Data underwrites every decision you make in your company. If your data isn’t reliable, your business outcomes will suffer. Here are just a few ways data quality promotes the success of your company:
Predictive analytics use historical data to predict future results. But if the historical data isn’t accurate, your projections will be off base as well.
Heavily regulated industries like healthcare, finance, and government operations could face millions of dollars in fines if their data isn’t complete and accurate. And as new protective regulations like GDPR place restrictions on the collection and use of personal data, all companies need reliable ways to manage information so they remain compliant.
The more accurate your data is, the less time employees will spend identifying and correcting errors or trying to find the information they need. Productivity gains contribute directly to your bottom line and that makes data quality an essential business priority.
Companies also rely on accurate data to make strategic business decisions. When data doesn’t tell the right story, it can lead to errors and missed opportunities. Good quality data reduces risk and promotes more advantageous outcomes.
Innovations like AI, IoT, automation, and machine learning all depend on data to work their magic. If the quality of your data is substandard, you won’t enjoy the full benefits of digital innovation.
6 Steps to Improving Your Data Quality
Bad data costs American companies $3.1 trillion every single year according to IBM. From customer service chatbots to industrial supply chain software, your company runs on data. Unfortunately, only 3 percent of companies have an acceptable data quality score, according to research published in Harvard Business Review.
If that doesn’t worry you, it should.
The good news is that as data volumes have increased, so have the resources available to help us manage our data. Let’s take a look at six ways you can improve your data quality to minimize risk and maximize opportunity.
1. Determine KPIs
You measure business performance based on indicators like cash flow, growth profit margin, sales, inventory turnover, and revenue growth. All of these keyperformance indicators, or KPIs, depend on data to give you the information you need to succeed. Before you can delve too deeply into data quality improvements, you need to identify the KPIs for each area of your company so you know what data to collect what level of detail you need, and how to create standards that ensure the integrity of your data.
2. Get to know your data
With your KPIs in place, the next step is documenting your data: what data you collect, why you collect it, and where it comes from. Many data warehouses are bursting at the seams with inconsequential pieces of information, and storing all that data gets cumbersome and/or expensive. A data audit will help you identify the data assets you own, determine how and where data is stored, and assess data usage and priorities.
3. Create data standards
Inconsistent data entry and formatting create huge problems for your data warehouse. That was the problem with the Mars Climate Orbiter. Don’t make that same mistake. Data standards ensure that data types and values remain consistent so you don’t end up with duplicate records or misaligned values. Processes like data matching, geocoding, data profiling, and data integration ensure that data adheres to format and entry standards.
It’s also important to document data elements, definitions, and validation rules. This is especially helpful if your data manager moves on to another organization and someone else must step in to lead your data initiatives.
4. Ensure data accuracy
Data validation identifies missing or inaccurate information and supports data integrity. It can be done automatically during the data entry process by flagging inconsistent or missing information. You can also validate data using techniques like loop back verification, which checks information against the data source.
Data accuracy includes normalizing formats, detecting duplicate entries, ensuring that your data sources are accurate and authoritative, and training employees on data entry standards. Data types and values should remain consistent across all records so that you can be confident in the accuracy of your reports.
5. Use a data warehouse to break down silos
You will never harness the power of your data until you can access it reliably. One of the most significant hindrances to effective data usage is storing data in multiple systems. These unconnected data silos cause problems with retrieval and analysis, and they make it nearly impossible to get a consistent picture of your business operations.
Sharing information across your organization helps you make better decisions, retrieve the information you need faster, and promote confidence in your analysis. That’s why data integration is so important. An integrated data solution promotes free exchange of data among all parts of your organization and supports quick, impactful decisions using the most up-to-date and relevant information.
6. Improve data structure
Data structuring is the process of organizing data so that it can be used. Ineffective data structures slow down your analysis and undermine the efficiency of data retrieval. There are many types of data structures, and using them strategically can prevent efficiency bottlenecks by putting data into a useable, manageable format.
Is all this effort really worth it? Think of data quality improvements as investments with exponential returns. Data management is a complex issue, and problems with quality and structure won’t be solved quickly. But when you begin making incremental improvements to the quality of your data you will see fewer errors, fewer costly mistakes, and more profitable business decisions. And that will prove to be worth it every time.