Executing data quality improvement projects is largely a matter of analysis, planning and learning how to work the system, say experts. In this section of the Data Quality Software Buyer's Guide, readers will learn why it's important to think like a senior-level executive when planning a data quality project. Readers will also get tips on how to keep data quality projects on track and on putting immediate business needs ahead of long-term data quality improvements.
Table of Contents
- Buyer’s Guide: Choosing data quality tools and software
- Gartner: Open source data quality software focuses on data profiling
- Data quality improvement projects require dollars and business sense
- Execute data quality improvement projects with senior-level thinking
- Gartner Magic Quadrant ranks data quality tools vendors
IT professionals seeking executive approval for data quality improvement projects should build a solid business case, come up with a plan and -- perhaps most importantly -- find the money themselves, according to data quality consultants.
While proper planning and execution is key to successful data quality improvement initiatives, such projects often languish for long periods of time when IT professionals fail to think like senior-level executives, said Richard Ordowich, a senior partner with Princeton, N.J.-based data governance consulting firm STS Associates Inc. He added that senior-level executives tend to think about money.
A former vice president of product development at Dow Jones Inc., Ordowich said that getting quick approval from senior management for data quality improvement and other IT projects is largely a matter of taking initiative with regard to budgetary concerns. He said it’s a critical step that many IT professionals below the level of chief information officer and chief technology officer never take.
“You’ve got to find the money,” Ordowich explained. “If you ask for the money, you’re going to spend months and years asking for the money. [Just] find the money, and then put your project together.”
One of the best ways to “find the money” is to approach someone in the finance department early on -- maybe even the chief financial officer -- to garner support for the data quality project, Ordowich said. Finance folks can be very resourceful when it comes to showing executive decision makers where to find room in the budget, he added.
“People call that politicking, [and] it is,” Ordowich said. “Well, I’m sorry, but that is the way society works. You can take the negative connotation out of that by saying, ‘I’m trying to get something done that I think is going to be beneficial to the organization, and I need support.’”
Keeping data quality improvement projects on track
Once the political aspect of the data quality improvement project is under control, it’s time to start planning and executing.
Data quality projects without the proper focus tend to fail or grow out of control, according to Danette McGilvray, president of Newark, Calif.-based Granite Falls Consulting Inc.
McGilvray, who is also the author of Executing Data Quality Projects: Ten Steps to Quality Data and Trusted Information, said words like business issues, goals, strategies or opportunities are worth repeating because they’ll guide the decision making process throughout the entire project.
“It’s really easy to have scope-creep, and there is never really enough time money or resources to do what we need to do,” McGilvray said. “Those words are going to be your guiding light to help keep you on track, so that at the end of the day, you’ll know you actually ended up spending time on the data that is pertinent to the [problem] you started out with.”
Putting the business ahead of data quality improvement
Ordowich, who has worked with several high-profile firms in the past, said the first step in executing a data quality improvement project is to conduct a “scan” of the organization.
A scan, which lasts about three to four weeks, is a preliminary analysis of just about every aspect of the company. The process is designed to ferret out the key data governance and data quality issues and identify their effect on the business.
“We look at the business, we look at the IT operations, we look at development, we look at their data processing and everything and try to get a sense of it,” Ordowich explained.
One of key the lessons Ordowich has learned from his scans -- a lesson that IT professionals should keep in mind when crafting their own data quality improvement strategies -- is that sometimes, in order to help the business, it’s necessary to fix the symptom of a data quality problem before addressing the root cause.
For example, Ordowich was once hired to help improve data quality at a company whose main function was selling data to customers via email. His scan of the organization revealed that there were in fact data quality problems, and customers were starting to notice. But the customers weren’t upset about the data quality issues, per se. What really irked them was that it would take weeks for the company to address individual data quality issues once they were reported.
Ordowich wanted to fix the customer-facing issue first because it had a direct effect on the bottom line. Oftentimes, he said, IT professionals fail to consider business-related factors like this when crafting data quality plans.
“I said, ‘Let’s fix your customer support [and] let’s get the fixes in faster so that the customers will be satisfied,’” Ordowich recalled. “Then we can spend time to fix the root causes of the problem.”
After identifying business problems and fixing customer-facing issues, the next step in executing a data quality project is to create a blueprint for fixing the data itself, Ordowich said. Problems that have a negative effect on the company’s revenue and business processes should be given the highest priority. But it’s also a good idea to focus on quick wins, he added.
For example, problems at the data-entry level can usually be fixed simply by educating personnel and making small changes to data-entry procedures.
The next step is to prioritize the big data quality problems that do require system redesigns and, finally, create a long-term roadmap for the future. A data quality improvement roadmap should cover the following two to three years and outline preventative measures that will keep data quality problems from resurfacing. The roadmap should also specify data quality objectives at the software development lifecycle level, Ordowich said. Doing so will help developers figure out which data quality tools and utilities to put into the design of their applications in the future.
“Developers typically are worried about writing code. They’re not focused on data,” the consultant explained. “The output of what they’re generating is data, but they don’t think about aspects like quality of the data when they’re designing their application. That has to be specified at the requirements level.”
Table of Contents
Guide: Choosing data quality tools and software
Gartner: Open source data quality software focuses on data profiling
Data quality improvement projects require dollars and business sense
Execute data quality improvement projects with senior-level thinking
Gartner Magic Quadrant ranks data quality tools vendors