August 24, 2019

The data is dirty

administrator No comments

The key to reliable transparency is always clean underlying data. And yet, as you progress, you will continually find data that is less than clean. In fact, sometimes it will be embarrassing. I’d argue this is a natural progression of the process. Let’s take a look at the problem and some potential solutions.


Damaged or incomplete data is usually a cultural problem first. You don’t have control of your data sources and there are no repercussions for getting it wrong.

 
Let’s start with the areas that should be reliable. Your GL data better be accurate. It’s audited every year to ensure just that. Even if there are small inaccuracies, you usually don’t have to worry. That’s the end of data that can be considered “clean” without work on your behalf. Depressing. 


Most TBM programs will need at least the following other sources.

 
A PPM system – this will create and maintain your project structure. Most likely it will house your time booking as well. Keeping this clean is just a meter of enforcement. Name and shame is the game here. If your business managers can’t keep this clean, get new ones. 


CMDB – You will map your infrastructure and other data attributes to your applications here. Many hands usually touch this system. The more hands, the more data issues. You will need to put application owners on the hook to ensure cleanliness. But the effort doesn’t stop there. Exception reporting, via Tableau or something similar, should become a part of the culture. Some errors, of course, can’t be picked up in reporting. Unfortunately you may find these errors in your allocation reporting and will need to be corrected retroactively. Keeping the CMDB clean is difficult but vital. 


Allocation database – this should house your allocation stats, percentages, and your other allocation driver information. Clearly it’s essential to the process. You have to limit access here. To many hands in the system causes big problems. I like to have only a couple people with edit rights here.


Those are the big three. Every other source enriches the process and has its own challenges. If you incorporate project management info into your reporting you have different challenges. Ensuring your PMs are consistent and accurate is a cultural undertaking. That goes back to my original point. You have to create a cultural that places importance on data. You can do that by fear, i.e. heads roll if there are too many issues. I don’t think that works long term. Reward folks who do the right things. And reward them in a public way. With enough focus on data, your TBM machine will run at a high level.