Windshield Wipers and Data Integrity
Tell me. When do you notice a problem with your car's windshield wipers?
Answer: When you need to use them. (And then, it's too late.)
On dry, sunny days, cruising along the highway, we never give a thought to our windshield wipers. We enjoy the ride. We get to our destination.
The same is true for data integrity. We don't spend much time thinking about our company's data or it's accuracy. We process transactions in our system. We serve customers. We generate reports, we close the books. 'Nuff said.
Then comes the day when we decide to upgrade our system, or do a deep dive into the data to solve a problem. And we discover that our data is terrible.
Just when we need to use it.
It's a lesson learned by EVERY company at least once in it's lifetime. Systems don't ensure data integrity, PEOPLE do. Sure, your system has edit checks, makes you complete all the fields and may even do some duplicate entry checking. But it doesn't do everything!
Anyone who has gone through a major systems implementation can tell you that data cleansing and data integrity of the legacy system chews up a significant amount of a project budget (as much as 10%-15% of the total cost).
And it's this crappy information you're basing business decisions on every day.
I've worked for a number of profitable companies who have learned this lesson. My last ERP project uncovered that within our legacy system, we were using more than 20,000 chart of accounts across the organization. More than 20,000!
Our Purchasing function had created more than 45,000 vendors in our legacy system. That's more than Wal-Mart! Even though the legacy system cost many millions of dollars, our Purchasing department couldn't prepare a report of how much we purchased from individual vendors. They asked the vendors to supply that information to us each year!
When we completed our project, our team had reduced the chart of accounts to less than 2,000 and our vendor list had been culled to less than a quarter of it's previous size. In the course of the exercise, we uncovered many duplicates, inaccuracies and data omissions.
Why? Because no one "owned" the data. Everyone relied on "the system". It's easy to blame something that can't argue back. Stupid software.
Usually, ERP implementations create new business roles. (note I didn't say I.T. roles). The folks who enter the data, who use the data, who rely upon the data, need to own the data.
In my case, our company created two new positions. Only one Financial person had the authority to create a new account within the system. This insured that no duplicate use accounts were entered because the data owner acted as the gatekeeper. You don't need THAT account, use this account instead (like everyone else). So few new accounts were created, and the size (and more importantly the integrity) of our accounting system remained intact.
On the Purchasing side, we dramatically reduced the number of people who were authorized to create new vendor records. This forced our purchasing users to spend a little more time searching for an existing vendor, rather than simply adding them again (with a slightly different spelling). Reports were created to monitor the growth of the account to warn us if new vendors were being created faster than expected.
Once the business took ownership for the data, the quality of the data improved dramatically.
The moral of the story?
Even though the sun is shining, take a look at your windshield wipers.