Making Sense of Big Data in the Real Business World
Executives understand the need for better quality, at least, intuitively, says Ted Friedman (@ted_friedman), vice president and distinguished analyst with Gartner’s Information Management team. It’s the math that trips them up.
During a Gartner webinar on improving data quality, Friedman said the most asked question was how do you measure the cost of data quality problems.
“People have a sense that poor-quality data is problematic for them, but I still continue to feel — and that webinar and the questions are another point of evidence on it — that most organizations have not done the math in a very rigorous way,” Friedman said.
There are many ways to measure data quality, and each organization needs to find the “personal” impact of data quality on its particular business strategies. Friedman outlined six basic ways to pinpoint the costs of poor data quality during a recent IT Business Edge interview:
1. The costs of reduced productivity, both for people and for business processes, such as when things take longer than they should because of data problems.
2. The costs of redundancies. If users don’t trust data, you wind up with shadow systems, where users are trying to “fix” the data on their own instead of relying on the enterprise systems. “You can actually observe business processes and observe people working and basically analyze the amount of time being spent working around and compensating for poor-quality data,” he said. “I’m giving you the simplified view here, but multiply that by the fully loaded cost of your labor force and there you go: You have an estimate of cost of poor data quality from an efficiency point of view.”
Another type of redundancy can happen with products. I’m familiar with a great example of this, where a manufacturer inadvertently overstocked the same o-ring, just because it didn’t know the o-ring was listed in its system under 200 different product IDs.
3. The costs of business processes breaking down because of data quality issues — which would include losing a customer because something went wrong due to bad data. For instance, a manufacturer may need to delay the production schedule if a data quality issue caused someone to under order or restock the wrong item.
4. The costs of compliance risks, such as when data quality problems expose you to regulatory violations.
5. The costs of classical financial risk, such as being unable to reliably understand how a company is performing due to data quality problems.
6. Lost opportunity costs. “Whatever growth might mean to any particular organization, I can begin to quantify those type things,” Friedman said.
Once you’ve identified the cost of data quality issue, you still have to find a way to fund it. I asked Friedman how data quality fared when it comes to competing with other budget goals.
“That’s why we advocate they don’t tee up data quality improvement as a competitive project against other things they may be doing, but rather attach it to those projects,” Friedman said. “Then articulate how and why those other initiatives that they’re trying to get funded may not succeed, could fail, or won’t achieve the benefits they need to achieve unless data quality is a part of them.”
In other words, he explained, data quality is a means to an end, not an end in and of itself.
“You don’t do data quality because data quality itself has some inherent value,” he said. “You do it because it enriches and makes more successful the business initiatives that you’re trying to run. So attaching it to those and making it an inherent part of those is a very valuable thing to do.”
