The IT industry needs to apply the quality management lessons already learned in other sectors since the 1920s, says Werner Swanepoel, senior manager at Deloitte Consulting.
One of these lessons is that achieving data quality involves organisational commitment and this is non-negotiable, as data quality management is not just an IT problem. It involves a thorough understanding of business processes.
"I don't want to disappoint anybody, but fixing data quality is going to take you long - years, not months," he said on the final day of ITWeb's BI Conference, in Bryanston, yesterday.
"You may fix your top five issues that will address 80% of the overall problem in the first year, but you're certainly not going to fix everything in year one. Especially large corporations - it takes two, three years before it becomes embedded in the culture of your organisation, and that's really where you want to be, otherwise you will consistently be fixing," he told delegates.
"If you're not serious about it, don't spend R5 million on the tools.
"We see many clients on a regular basis where they tell us, 'Oh, we have the tools. We've got different tools because when this one didn't work we bought the next one, and it still hasn't fixed the data quality problem.'"
Headache pill
Swanepoel said there was a link between data quality and business processes. "It is fundamentally true. Nine out of 10 data quality issues are related to a business process that's not functioning properly."
He added that identifying the root causes of data quality issues was essential to addressing the matter.
"If you go to a doctor with a headache, you would like him to check up slightly more than just give you a headache pill. But we see it happening all the time in organisations, where they just take headache pills. They never spend the time to understand what the root cause is."
Swanepoel says quality control is not unique to IT, and much can be gained from the lessons other industries, particularly manufacturing, have already learned. Formal quality control principles have been in place in the manufacturing sector at least since the 1920s and modern-day interpretations of total quality management still use tools derived from those early days.
"The challenge is for the IT industry to embrace some of these concepts, some of these techniques, go back to what has been done and what has been refined in manufacturing production for many years and understand how you apply some of these techniques when you try to manage data quality."
Related stories:
Locals 'generally happy' with BI
BI to the rescue
BI consolidation a good thing?
Share