Did you know that before Schrödinger shoved a cat into a box he was experimenting with putting spreadsheets to file shares and creating client records in CRM systems?
I know, I couldn’t believe it either.
So what did he find out?
In Schrödinger’s late 20s, when he worked as first Data Management Consultant, he realised that he could never confidently say if data in any system is correct until he looked in. And it bothered him because he saw it over and over again, so he started to ponder what was behind that paradox.
I won’t bother you with details. Not because I am making this up in a spur of the moment, but because his research was incredibly detailed and in-depth, so I’m giving you just a summary.
Let’s transport ourselves into a world of retail sales. ZHhhooooooooooom.
When you work in an operational role where you need to make sure that everything is done and it is done the right way. Let’s say that you process and fulfil new orders. Every day you make sure that you pack the right things into the box, pack it properly, slap an address on it and ship it. Which orders and customers are important at each moment? Most likely only the orders in progress and the ones which are getting returned. I hope no one falls asleep thinking about an order they completed two years ago. Even worse, thinking about all the ones which have been ever been done!
It is the same with data since it’s just a footprint of some activity. It is natural to think about what is in our “to do pipeline” and not about all the history and any customer ever. It’s just incredibly difficult to keep it in your head all the time otherwise.
So what happens if an order is completed within a week on average? How long does it stay relevant and important?
For a person packing the box, it might be one day?
For the fulfilment team, it might be a week?
Team processing returns will keep an eye on them for 30 days.
Finance team managing payments and their reconciliation for 60 days?
The financial planning team and product team might be doing forecasts based on the last three years?
When an acquisition comes there might be a need for the whole history, so it’s relevant forever?
You might notice that further down in the list you go, you will need current order data less frequently. You might check an order daily in the first two days, then check it at the end of the week that it has been received, reconcile payment after a month and then?
So what happens when you look into a system? Is data correct or not?
And why is it important? You might notice that usually, it’s different teams that will need the data later. However, they don’t know anything about the state of data, systems or processes until they look. How could they know? They are not keeping in their heads that there were system migrations, upgrades, changes into both systems and processes. Not even talking about bugs, accidental mistakes or workarounds.
And what if they don’t even look, but their work is automated? Unless it’s done by Bender from Futurama, the automation script won’t know if data is off. Most likely it won’t be “we don’t have a status field populated for any order” but it will be just enough to annoy someone that you don’t meet sales target because 5% of orders in Q1 2021 have status field blank.
If we want to get more value from data either to get some insight or to automate a process, we need to move from “I will look and then see if data is correct”, to be able to just do it without even thinking about it. As data and system owners, we need to start living in a world where we will just be visited at any point in time and we need to make sure we have our house tidied up. As an organisation, we need to understand that it is a lot of hard work, which was assumed that has always been done, but rarely was. Mostly because there was never any time to do it! We are also increasing the number of dependencies we didn’t need to deal with. We might not even have the proper tools and skills to manage those dependencies the right way.
It’s pretty much like the famous song “My customer data brings all the boys to the yard”.
And there is a quick test you can do.
How much of your MI which goes to your executive team, board, regulatory bodies, out to the public or is used for any incentives is fully automated? How much of it is put together manually, because it has to be cleaned and corrected every time it is done?
Just above average 