Your business outcomes define what data quality is for your organization

Your business outcomes define what data quality is for your organization

Whenever the conversation of data quality comes about, often it is tied to whether the data is accurate or not. However, data quality is more than just the accuracy of the data. Data quality can be defined as a measure of how well-suited a data set is to serve its specific outcomes. So, data can be accurate and yet not be well suited to help the business objective in question.


With this definition in mind, we realize that there are other metrics we need to use to determine whether we have quality data or not. These core metrics are Accuracy, completeness, latency, consistency, and reasonableness. Each of the metrics may differ in their definition and importance depending on the business outcome in question. As such it becomes imperative to have the business outcome clearly defined since these metrics rely on the outcome.


1. Accuracy: refers to how correct the data value is in comparison to a reference source. In this instance, the correctness of the data needs to be verifiable with some source. The degree of accuracy may differ depending on the business outcome. For instance, the finance department may require the financial data they work with to be 100% accurate. However, in the case of working with third-party data say customer sales data or even customer online traffic, the true accuracy of the data becomes difficult, if not impossible to determine. However, in this instance often are looking for trends over a given period and the accuracy of the data cannot be 100% guaranteed.

2. Completeness: Completeness is based on the user’s needs. Does the data offer complete information that would allow the user to perform the required task? If an analyst needs to analyze all the company’s products by store, they need to ensure that the data they have has all the information required for said task.


3. Latency/Timeliness: This refers to the time between when the real-world occurrence happens and when it is available in the data. Some customer sales systems update weekly, this would be a problem for you if the company has just launched a new product and you need to track daily sales. The latency period of the data is then defined by the business need. If the specific task requires real-time data, such as the stock price, relying on a third-party site that cannot deliver would not be prudent. However, if you are tracking a digital marketing campaign then real-time data is not needed.


4. Consistency: Consistency refers to the uniformity of the data across the various data phases it may pass through. The values of the data should remain consistent. Furthermore, data should align in meaning and format and provide a single picture of the business outcome in question. If the organization is selling your products through large retailers and have to rely on their customer sales systems, then when the data is extracted, it is necessary to adequately unify the data to maintain consistency within the reporting processes of the organization.


5. Reasonableness: Speaks to the validity of the data. Does the data accurately relate to the business domain it seeks to model? This is a question on the credibility of the data. It is a proper representation of the real-world instance in question. If most of the company’s customer base is 70% female, then having a data set that has 60% males would not be a reasonable depiction of your customer base.


The question of data quality is a rather complex one. The complexity is eased with clarity on the business outcome in question. Once the outcome is clear then the quality of data required to speak to this specific outcome can then be defined. As more and more organizations move towards leveraging data to drive their business objectives, data quality will become more imperative not only to define but to maintain. It is the organization with high quality to data that can be used to derive meaningful business levers to pull that will have the upper hand.

To view or add a comment, sign in

Insights from the community

Others also viewed

Explore topics