Data visualisation: following the money
19 March 2018
I was talking to a friend recently who has an IT system where she works that gives things away free. It doesn’t mean to, it just lets the users close out contracts without going through the manual step of raising an invoice. When the friend had looked into this, there were dozens of missed invoices totalling a five-figure sum stretching back over three years.
Fast forward a month and I was running some business process visualisation for a client, pushing all of their data through a tool that draws out the pathway through each business process. What should emerge, but a little red line that connected the delivery of services straight to the contract close, missing out the invoicing stage. In this case, the missed billing just nudged into the six-figure category.
So what was going wrong?
In the first case, the application was lacking in functionality and required significant manual intervention to keep contracts moving. Each human touchpoint introduced scope for missing information, miscalculation, or skipping required steps in the process. In the second case, the integration of multiple complex systems had led to an architecture that no one really understood. They knew it worked most of the time, but they had no idea where all the cracks were that let processes stall or fail.
It seems hard to believe that these things would really happen, but they do. That is why I’m such a big fan of data visualisation. When you take every transaction in a system and look at how they move it becomes easy to identify the outliers and unusual events. Not only can you see missed invoicing, but also you can see where approvals are bypassed, where mandatory steps have been circumvented, or where transactions have just got plain old stuck.
It’s a huge eye-opener to see your data in this way, gazing into what can otherwise be a black box where you cross your fingers and hope it always works. For me, it should be a component of the future audit, the backbone of many forensic investigations, and a way of checking up on major business systems implementations. What’s more, that only covers the risk and controls angle. The best tools can show how optimised you are relative to your “happy path” (the ideal process that transactions should follow), timings between steps in the process, and segmented analysis of each process to help you understand situations where the process works best.
There are so many different possibilities for using data visualisation technology to optimise processes but as these two cases show, sometimes you just need to start by getting the basics right.