Back To Blog

Top Five Reasons Utility Data Management & Analytic Projects Can Fail

Share on FacebookShare on Google+Tweet about this on TwitterShare on LinkedIn

 

Bringing the power of analytics to a complex organization, like a utility, is a challenging endeavor. It requires a lot of bright people with a detailed technical understanding of the issues and their causes. You would think that if you can just find the right analytic platform, the best learning algorithms and the smartest subject matter experts, you could revolutionize a utility, but there’s more to it. While change is possible, the reality is that organizations are failing left-and-right at these projects. The reasons they fail may not be the reasons you might think. From experience, we’ve found most failures can be distilled down to the following five causes.

 

1. Goal? What was my goal?

Most data management and analytics projects start with reasonably clear goals. These endeavors often start with a list of problems that need to be solved, such as being able to access more data in one application or provide insight into the health of a certain class of asset. As more and more stakeholders get involved, politics and new ideas can start to expand the scope of the project. As the scope grows and grows, the chance of success shrinks and shrinks. Once the goal of the project is lost, so is the project itself — it’s hard to get to where you want to go if you’re chasing a moving target, or if the team is not pulling in the same direction.

TIP: Create a clear business goal for the project along with an outline of the decisions the project will support, and stick to the plan.

2. Data is messy.

Don’t trust anyone who tells you their data is in good shape. It’s safe to assume most data is lost, sparse, and in the wrong format. It’s also common for data to be found throughout the project. Like your initial information, this data will need to be cleaned up, which will add more time to your schedule. Don’t assume you can just hire “data guys” to clean it up and do the work for you. Successful data cleanup requires getting the appropriate subject matter experts to review the insights to make sure the cleanup process is on track and that critical findings aren’t lost along the way. This step is one of the costliest parts of a data-related project and can often eclipse the cost of analytics and data management platforms. If you don’t budget for this, your project will fail.

TIP: Reach out to other utilities that have or are going through data cleanup activities. Without insights from experience, the cleanup process will be a difficult task to fully appreciate.

3. Data modeling is hard.

All the data you’re collecting, cleaning, organizing and analyzing needs to go somewhere. What form should that data take, what structure should you follow, and which standards need to be applied? These are just some of the questions you need to answer. Often, too much time is spent on the aesthetics of dashboards and finals displays that the data structure requirements are an afterthought. The result is a “data swamp” of competing formats. There are a variety of models available through common information model (CIM) that should be considered. It’s also important to remember good, old-fashioned data management best practices.

TIP: Learn about and fully understand the implications of IEC 61970, 619268 and 62325 for CIM and energy management. You may want to participate or reach out to the CIM User Group as they are a very active group with knowledgeable people.

4. Cybersecurity and contingency planning takes a long time.

Cybersecurity and contingency planning are on everyone’s mind. We all know it’s important, but groups often wait too long into the project to begin auditing and reviewing security infrastructure. Cybersecurity is about first installing the appropriate security software, testing it, validating it, and then making corrections. Often a review of source code is required along with dedicated intrusion testing. It takes several times longer than anyone usually believes.

TIP: Start cybersecurity and contingency planning early in the project and leave plenty of time during the process to conduct proper intrusion and contingency tests. The project isn’t over until the cybersecurity and contingency plan testing is complete.

5. Who’s the boss?

No upper management involvement is probably the most common reason analytics and data management projects fail. Data-related projects span multiple departments, and all stakeholders have different needs and requirements. Someone needs to be able to put down the rules and own the budget. That person needs to own the budget of multiple departments and make it part of the entire company’s goals. This person must also be present at key meetings and make sure everyone is making progress and mitigate the politics.

TIP: Depending on the size of your organization, this “big boss” could be the President, CEO, Senior Vice President or Chief Technology Officer.

Data is undoubtedly king in the increasingly digital world in which we live, but managing critical information successfully can be tricky without the right approach. Customize the guiding principles we’ve described to fit the business’s unique needs and with a little patience and teamwork, your team will be well on its way to effective data management.

Additional Information

Back To Blog

Categories

Authors

Tags

Follow Us

Subscribe to our Blog

  • Enter your email address
  • This field is for validation purposes and should be left unchanged.

Doble Engineering Company

Doble Engineering Company