6 Steps to Fix Bad Data Issues That Cost You Money

Improper data management can cost a lot of money. From administrative fines to extraordinary expenses to obtain accurate information to make critical business decisions, Y It may even be necessary to incur expenses to clean a bad image due to a data error..

Most organizations tend to handle data quality in a very tactical way. Either the IT department solves any problem or the different business units must create business processes to solve it.

When IT has to take care, the problem is queued and, often, low priority is assigned. Therefore, business units, who face the problem on a day-to-day basis, they must intervene.

Later, Why are data quality issues not addressed at the organizational level?

In this article, We will describe six simple but very effective steps to prevent and correct erroneous data..

How to troubleshoot bad data

1. Understand the problem

The organization must reach a consensus on where the bad data is and what impact it has on the business.

The best way to do this is by profiling data.. Data profiling is a learning activity, focused on finding out what it has and how bad it is. Different types of disparities can be discovered during this profiling process.

Data may be incomplete. For instance, customer records in a customer relationship management system (CRM) may lack zip codes or email addresses, or the part numbers may not appear in the entries of a materials management database. . O, an HR app. may not have employee records. In this second case, you must determine the full definition of employee to perform your assessment (namely, if for example contractors should be included and / or subcontractors)

You can also discover inaccuracies or inconsistencies within the data.. Are the values ​​contained in the database fields correct?? Some common problems may include postal code entries that contain letters or customer email addresses without the symbol “@”.

Finally, duplications and layoffs are likely. For instance, products may have entered an inventory database more than once, or there can be multiple records for a single customer in a CRM system.

2. Create the figure"Figure" is a term that is used in various contexts, From art to anatomy. In the artistic field, refers to the representation of human or animal forms in sculptures and paintings. In anatomy, designates the shape and structure of the body. What's more, in mathematics, "figure" it is related to geometric shapes. Its versatility makes it a fundamental concept in multiple disciplines.... of the data controller

The data officer may be the most important person in your data quality initiative.. He will be responsible for creating the rules for the generation, management, data maintenance and distribution, as well as outlining the processes that will help ensure quality throughout the organization.

The person responsible for the treatment also Ensure compliance with these guidelines, continuously monitoring and measuring the integrity of information and modifying quality practices as needs change., data sources and other factors.

3. Accept the impact of bad data

You must get one comprehensive view of the impact of Bad Data. Through the data profile made in the first step, you come to understand what is wrong. In this step, determine why it is wrong and how it is affecting your business.

Perhaps the best way to start this step is to create a lifecycle chart of the data that was observed during the profiling step.. Where does the data come from? What are the downstream applications that use this data? What is the impact on the following applications if the data is not correct?

Identify touch points during the lifecycle where data can be manipulated and highlight places where data is produced, so you can clearly see what happens when quality is compromised.. In every step, associates a value with good data and bad data. When the life cycle is complete, just add the numbers.

4. Decide what to do with the wrong data

Once problems are discovered during profiling and you have an idea of ​​the magnitude of the problems, What will you do to correct them? All opportunities must be prioritized, challenges and risks, and it should be delineated, implement and apply a methodology to manage them.

The plan must answer these questions:

• What errors will be accepted (those within acceptable limits) and which ones will be rejected (the most troublesome mistakes)?

• What will happen to the rejected mistakes? Will the data be excluded or will it be sent to someone for further review?

• How will the correctable errors be corrected?? Will they be handled manually or will defaults be applied?

5. Start the cleaning process

Once the problematic data is identified, how they should be corrected is determined and responsible persons are assigned, Whats Next?

Time to implement the processes, procedures and plans to start cleaning the data.

This may seem like a technical exercise and should involve your IT organization working with its own tools.. But Following the steps above will make applying the plans a breeze.

6. Implement and maintain general scope data quality processes.

You now have a process to solve the initial business problem and it is working more efficiently. Does it work most effectively? Are there other data-related issues that your business unit should look for to correct?? If so, has the people, the processes and technology to deal with them.

But nevertheless, Before I start talking about it, you need to make sure you have some way to measure and monitor progress towards your data quality goals.

The metrics have already been developed and it is important to keep them continually updated.