Proper data management is key in any Big Data or Data Lake, analytics projects. Here are my top 5 rules for proper enterprise data management for analytics.
- Understand Your Data – I think this goes back to the first few weeks of class where we talked about sorting or organizing your data. It is impossible to move on if you are unsure where to start. You have to understand what is right in front of you if you are to create an action plan.
- Drive Outcomes, Not Insight – We have a saying in sales, is what you are doing a revenue generating event? This is pretty much the guiding stick to everything we do. If it not helping a sale then it is unnecessary. Everything we do in our professions have to have a purpose. This is the case with data analytics as well. Insight can drive outcomes but it’s not a guarantee.
- Automate everything non-valued add – This refers to my previous comment about driving outcomes. If possible, you should eliminate unnecessary tasks to free up time. Time is a valuable commodity, you are losing it every day and there is no way to get it back. The best example of this is an article I read a while back with an IT guy who automated every part of his job that took over 90 seconds to complete. After the guy left for a new job, his former coworkers were looking through his work and discovered that the guy had automated all sorts of crazy things, including parts of his job, his relationships, and making coffee. This maybe taking it a little too far but he has the right idea.
- Minimize Custom Development – My manager told me a story once about when he worked in IT for EMC. He said EMC made so many tweaks and changes to our SAP application that SAP would no longer support it. One day it broke and when they called SAP to fix it they said “that is not our application anymore”. Sometimes changes make things more complex. It’s best to have a standardized approach to help minimize issues and errors.
- Keep Things Simple – Though not always possible, it is best to keep things simple and straight forward when you can. This goes into the time management piece of a lot of my points. If you can make something simple it will most likely be repeatable on a large scale. This increases efficiencies by decreasing time used and the amount of potential errors.