Data Quality Is Not the “Problem”

Data quality is not the “problem”. I made the point in a discussion in my digital transformation class. Let me answer some questions about my claim.

• Why is data quality not the “problem”?
• What are the real problems?
• How to solve the problems?

Data quality: Let there be light and order (An image generated by OpenAI)

Why is data quality not the “problem”?

I took a peek at your iPhone contact list. There are duplicate entries, missing information, outdated numbers, and messy formatting. But I do not regularly clean up my contact list? Do you? Now, let’s consider data quality in the business context.

First, there is no universal definition of data quality. Let’s use customer information as an example. For marketing purposes, an email address is good enough to create awareness of our products. For salespeople, a first name and a phone number are a good starting point to build a relationship. When it comes to order fulfillment, we need a shipping address. As for accounting purposes, I’d better stop, as an accountant once told me: “Accounting makes you cry.”

Second, we do not need perfect data. As strangers become prospects and then customers, they gradually share information with us. We earn their trust to get just enough information to drive the business forward.

Third, data quality can take a lot of investments.

• Expertise
• Time to do data collection, analysis, transformation reporting
• Tools and technologies

Such investments need business justification.

The problem of business is to create value for stakeholders, including customers, employees, investors, and communities. While data quality is a critical factor, it is a means to an end and not an end in itself. Therefore, data quality is not the “problem”.

What are the real problems?

Let’s look at some examples. Each can lead to a significant business case.

• We allow anyone to try our SAAS (Software as a Service) product. But how do we know who is likely going to convert to a paying customer?
• How do we know whether our customers are happily using our services?
• Can we proactively detect and fix problems for our customers?
• Our executives eagerly wait for reports to make strategic decisions while our analytics team is doing a lot of manual data manipulation.
• We need to be SOX compliant.

With these specific problems, we can define data quality, then invest to achieve the business purposes.

How to solve business problems with data quality improvement?

To be successful with data quality, we need quite a lot —

• Executive sponsorship
• Clearly defined goals and scope
• Stakeholder engagement
• Project team with data expertise
• Data quality rules
• Ongoing data governance
• Continuous improvement after initial implementation.

In conclusion, data quality is a critical factor in business success.. However, it is not the real “problem”. The real problems lie in identifying and addressing specific business needs, such as customer acquisition, satisfaction, and retention, reporting and compliance. By addressing specific business needs, we can achieve just good enough data quality just in time.