Six dimensions of data quality

As every company becomes a data company, improving the lifeblood is of pivotal importance to remain as an important player in the market. Data quality is measuring the condition of data. In this way it ensures that the data of your organization fits the purpose. It also helps organizations to identify data errors which need to be resolved. It’s important to have trustworthy data where you can rely on to make important business decisions.  

Six dimensions of data quality 

Okay, this is the theory. That’s clear. How can you make a good distinction between good and bad data? There are 6 dimensions of data quality which can be measured.  

• Consistency: Data values from different databases should match or align with each other. This means that data has no contradictions in your databases. For example, if you look at your R&D-budget in the company, this should be in every database the same and can’t exceed the total budget for that year.
• Accuracy: The data is correct and exact. A measured value should match with the actual value and has no errors in it. 
• Completeness: There are no missing or incomplete values in the dataset. If you want to have a high quality data set all data entries must be complete. 
• Validity: Data validity ensures correctness and reasonableness of data. This means that all data points should follow the same and correct rules. So, for example, rounding percentages to a whole number or formatting dates all in the same way. If at some data point other rules are used, then this is an error.  
• Uniqueness: Data will not be recorded more than once within the data set. However, you can use the same data point in multiple ways, such as costs appearing in both a management and sales report. But you can’t duplicate the data points, by listing twice the same number in a certain report.  
• Timeliness: Data is available, accurate and is updated as frequently as necessary. You should look at your data and see how often you should track the data. Does the data have a short-term impact, then it’s better to measure the data on a monthly basis instead of annually.  

If the data of your organization fits all these dimensions, you can speak about good data quality. All these dimensions are important to measure data quality, however some organizations should focus more on one dimension than others to support specific main goals of your organization. 

The benefits 

Now we know what data quality is and how you can measure it, however why should we bother so much about good data quality? Good data has a lot of benefits.  

First of all, it can help organizations to make better business decisions. When the data is of good quality, then it is easier to process and analyze the data. This leads to better insights for the organization and easier decision making.  

Secondly, it leads to more productivity and efficiency. If all the required information is all in one quality dataset, teams can directly see all the information they need. If they must work with the raw unchecked data, then they have to check it every time themselves and open multiple datasets. Checking all the incorrect or missing values costs a lot of time, so when teams can just open one very good data set the efficiency can increase dramatically.  

Thirdly, the cost of bad data is very high. There has been multiple researches about this topic, which also indicates the importance of data quality. IBM once calculated that the annual cost of data quality issues in the U.S. was 3.1$ trillion. Other research from a.o. MIT Sloan found that business dealing with the consequences of bad data costs a company around 15%-25% of their annual revenue.  

What about you? 

You might wonder, which are those errors and why are the costs so high? Some examples of what bad data can cause are added costs when the shipment of certain products goes wrong. If the wrong product is sent to a customer, you first need to make sure the right order will be delivered and most of the time the customer also gets a discount coupon to make up for the mistakes. Moreover, the company can get fines if their financial statements are not properly reported. And when you don’t have the right customer data, you might target the wrong customer group and this can lead to less sales then when you had targeted the right group. If you want to have insights about your data quality, why not meet up with an expert data doctor who can help you and make a first diagnosis? We are ready to help you. Human Inference, more than 35 years experience at your doorstep!