Why CXOs ignore data quality problems

Too many CXOs ignore data quality problems until they result in catastrophe. Here's why, and how to avoid the mayhem.

Image: iStock/DeanDrobot

There is virtually no business without a data quality problem somewhere. It might be an incomplete record of a service call, an erroneous name or address spelling of a customer, or even an errant billing. When data quality issues like these arise, they are usually the result of human error in systems of record that have operated for many years. C-level executives and line managers tolerate them, and in most cases, the data quality problems are simply worked around and forgotten.

Now with the influx of vast piles of unstructured big data, these historical data quality problems have multiplied exponentially--but the C-level response of relative indifference remains. Should it?

As a big data industry analyst and commentator, the logical answer for me is no. However, there are there are actually myriad reasons why CXOs feel justified ignoring data quality issues.

To start, there are the ways in which CXOs see big data. Typically their view is through dashboard eyeshots of how their business operations and/or sales are going. It's also through the milestones achieved in new big data initiatives undertaken by their staff, that CXOs must report to their boards.

In a nutshell, these C-level execs are looking for project completion milestones and results. They aren't focusing on whether their data analysts are spending 75 percent of their time cleaning up data to prepare it for analytics. And a new generation of CXOs who have grown up with Internet and rapid app prototyping are more than comfortable if an app doesn't work correctly 100 percent of the time--as long as it basically works.

Another one of the reasons data quality gets ignored is because it lacks urgency. CXOs are asked to keep the company profitable, to build the company's brand, and to return benefits to stakeholders. CXO performance evaluations, compensation and boardroom conversations revolve around this. Along with these obligations come all of the daily responsibilities that CXOs must address. If equipment breaks on the manufacturing floor, or if safety conditions at a company facility are challenged by regulators, CXOs must take charge. If a major customer is disappointed at a botched order, the CEO might have to get on a plane and personally visit the customer to get the situation back on track. It is hard for CXOs to worry about data quality when these major daily exigencies must come first.

SEE: Data's new home: Your company's balance sheet

What happens is that big data (and other data) initiatives move forward until the flawed data prevents migration to a new system, or something catastrophic happens, like a data breach or an errant forecast produced by poor quality data. An example of this was the faulty Google flu forecast, which was based on errant data and assumptions.

When data miscues like this happen (and Google is not alone in this), that's what grabs the attention of C-level executives because their boards and their stakeholders are there demanding answers.

In a recent visit I had with Dan Ortega, vice president of marketing for data intelligence and accuracy solutions company Blazent, Ortega said that improvement in company data initiatives must "come from the top down, from the level of the CEO or the general manager."

He's right. The best thing that CXOs can do is to avoid getting into these dire situations. They need to acknowledge that their already hectic schedules cannot absorb additional attention to issues like data quality so they can find ways to avoid embarrassing situations that can compromise corporate revenues, performance, or brands.

One proactive step that can be taken is to assign data quality to the corporate risk management function that already shocks loan portfolios in financial services firms (to check them for stability), weighs the possibilities of supply chain disruptions for shippers and retailers, or evaluates the results of potential economic downturns and/or regulatory and product failures. Risk management is a logical place to add data quality--and it is likely to become even more important if data one day becomes a tangible asset that is valued on corporate balance sheets.

Will elevating the stature of data quality by placing it within the corporate risk management function solve every data quality issue? Of course not. But by making data quality a new risk management category, it will assist CXOs and those who work for them in developing a methodology where data quality can be kept under the scope in the same way that earnings, revenue generation, operational costs and brand performance are.

Also see