By Elliot King
The real question is not whether data matches some pre-established and perhaps arbitrary mark. Instead, perhaps the most important challenge for data quality professionals is to determine if the quality of the data meets the standard needed to complete the task at hand. And the only way to know the answer to that question is to ask. Remember, IT does not own corporate data. The business users do.
With that mind, a regular, ongoing program to query key consumers of data should be integrated into the overall data quality initiative. Data consumers should be asked if they can access the data they need efficiently and easily. Is the data they access timely? Is the data accurate? Is the data relevant? And is the data consistent?
That feedback can be used to identify and address pressing data quality needs. In many companies, different business units can tolerate poorer data quality and still complete their tasks at hand--or so they think.
Front-line data consumers are only one constituency of interest, however. The person responsible for the data in a business unit must be queried as well about data quality issues. Front-line users may feel satisfied with the level of data quality, but a manager with a broader perspective may feel otherwise and vice versa.
Finally, senior decision-makers must be brought into the conversation. Effective decision-support represents one of the most potent potential payoffs for investment in data quality and also represents one of the most devastating potential pitfalls. The data quality stakes in decision-support can be very high.
Data quality improvement programs can be laborious and costly. To be effective, they must be crafted to identify, target and address the areas in which poor data quality has the biggest impact on company operations. To identify those weaknesses, data quality specialists must ask their customers, the users of the data.
Here are some simple truths that too many companies ignore. Concerns about data quality cannot be confined to the IT department. No purely technical solution can insure high quality data. And, oh yes, no matter what you do technically, data will never be perfect.
The real question is not whether data matches some pre-established and perhaps arbitrary mark. Instead, perhaps the most important challenge for data quality professionals is to determine if the quality of the data meets the standard needed to complete the task at hand. And the only way to know the answer to that question is to ask. Remember, IT does not own corporate data. The business users do.
With that mind, a regular, ongoing program to query key consumers of data should be integrated into the overall data quality initiative. Data consumers should be asked if they can access the data they need efficiently and easily. Is the data they access timely? Is the data accurate? Is the data relevant? And is the data consistent?
That feedback can be used to identify and address pressing data quality needs. In many companies, different business units can tolerate poorer data quality and still complete their tasks at hand--or so they think.
Front-line data consumers are only one constituency of interest, however. The person responsible for the data in a business unit must be queried as well about data quality issues. Front-line users may feel satisfied with the level of data quality, but a manager with a broader perspective may feel otherwise and vice versa.
Finally, senior decision-makers must be brought into the conversation. Effective decision-support represents one of the most potent potential payoffs for investment in data quality and also represents one of the most devastating potential pitfalls. The data quality stakes in decision-support can be very high.
Data quality improvement programs can be laborious and costly. To be effective, they must be crafted to identify, target and address the areas in which poor data quality has the biggest impact on company operations. To identify those weaknesses, data quality specialists must ask their customers, the users of the data.
Leave a comment