Rick and I go back few years. He was in the Head Office as head of quality program and I was the IT person supporting him. Then 3 years back, he went on to become plant manager.
This morning, he stepped into my office and announced that all the reports I had given CFO, were wrong. He went on to tell me that the numbers I reported for his plant for last month were completely wrong. I was a bit surprised and puzzled first. I asked Rick to explain why he thinks that the reports were wrong.
He turned to page 3 and showed me that the person who worked most in his last month did not exist. In other words, someone by the name of Ron Clarck who found most of the problems while doing QA, did not exist or rather does not work at the plant. I was a bit surprised but hiding my surprise, I asked him if he found any problems with rest of the matrices about % defects, man hours of rework, Delayed shipments due to defects etc? He sheepishly accepted that there was no problem with those numbers but he continued to argue about the quality of report that includes name of a person that does not work at the plan.
I asked Rick if he would like to have lunch with me? He thought I was trying to bribe him but since we are good friends, he agreed.
During lunch, I talked to Rick about Data Governance model we have adopted in the company.
I explained to him that as a business, we had to prioritize the pieces of information that had to be right and then there were pieces of information that we as a business could afford to have wrong as long as those were within manageable %.
Rick responded by saying that as far as he was concerned, data is either right or wrong. It can’t be both. Rick lives in 2 colored world. In finance, we recognize that there are shades of grey. But he was inquisitive enough to ask how we “prioritized” data. I remember him using quotes with his hands to suggest that he does not believe that it is possible. He had heard of data cleansing activity being undertaken by organization but data governance was new to him. Since he had given me an opening to explain Data Governance, I decided that my burger will have to wait for a while. While explaining to him, I started working on my French fries.
I started the explanation by first explaining the process to him. The process we followed was to get data fields from all the systems that are in use in our office. I had one of my interns identify the duplicate fields from multiple systems. We first attacked finance related fields first. While getting the data fields, I also asked them to give us the data types and the field lengths allowed. Once we had an exhaustive list of fields, we first identified the fields that had different definitions in various systems. Along with data fields, we also got some narration of the fields. In other words, we had data dictionary for the organization with us.
Interestingly, we found one field which had 2 different lengths in the same system.
Once this exercise was over, I asked my team to identify the fields that would have impact on key business indicators that were in focus for the year. At the beginning of the year, our organization works on identify key indicators that will be in focus for the year. Examples of these are day to first shipment, First shipment fill rate, Promises kept, requests met etc. This was not a simple exercise. It had to be done by breaking the business indicator down in multiple layers till data fields were identified. For example, Days to First Shipment, was an indicator that does not have bearing on the name of the person who conducted the final QC. But that indicator gets affected by the date of shipment, customer name, customer address etc.
Once this list was made, we had the list of fields whose quality had to be very good. Very good meaning 98% or better.
All the critical fields were then analyzed and each system made a presentation on how their system either using screen control, business rule or database definition, ensures better quality. As part of this exercise, we identified many best practices which were recommended to all the systems.
You guessed it… QA person’s name was not in the list of critical fields. Ron Clarck was really Adam Ronald Clark. Since he was going by the name of Adam Clark, no one knew that in IT system of the plant, he was defined as Ron Clark. The QA system used to pick up data from Single Sign On record.
Mystery solved. Thankfully, we did not have ghosts working Quality Checking our products.
Not sure what Rick thought about our lunch but I was convinced that the months of hard work that my team put in to improve data quality was starting to pay dividends. People were paying attention to the reports. Now the “system” was telling them how they were performing. They could ask questions but point fingers was quite futile.
Rick paid for my lunch. I wonder why….. 🙂
One critical aspect I want to mention here is that data governance initiative is a business initiative and has to be managed by the business. IT can do bulk of the data collection, reconciliation and identification of problem areas but governance has to be driven by the business.