Outsourcing and Analytics: Foresight Makes the Difference
Outsourcing has emerged as a way to support utility analytics activities, but it comes with trade-offs.
Outsourcing IT and business processes is not new to utilities. With the increased popularity of analytics, outsourcing has emerged as a way to support utility analytics activities. Like everything else, though, outsourcing comes with trade-offs. Having the right contractual relationships in place, along with processes and skilled management, will determine the difference between success and failure.
A little history
In the mid-2000s, utilities that outsourced did so to support functions such as application/infrastructure management, credit/collections and systems integration. Later, some opted to outsource application/web development and IT maintenance/support. Fast forward to the analytics era. One utility analytics leader says, “Cloud infrastructure and data management are disrupting traditional categories.” Utilities can outsource for data hosting, analytics processing and data science expertise–or all of the above through cloud-based analytics platforms.
The most popular offerings are specific to utility industry business areas. Examples of analytics platforms are: GE with asset performance management, Bidgely with customer analytics and IBM with vegetation management. Vendors that offer “data science as a service” offer skills as well as platforms. For example, Trove, which recently was acquired by E Source, offers “data scientists with strong domain expertise to augment” a utility’s analytics team. For more generic analytics functions, cloud-based platforms such as AWS and MS Azure are being used, with utility-specific analytics layered on by other vendors or in-house teams.
Outsource – or not
Outsourcing is quicker and maybe less expensive than standing up in-house infrastructure. Utilities that are just starting to explore analytics face a chicken-and-egg situation. In-house skills or infrastructure are lacking, so it’s not easy to make a case for a corporate analytics strategy. In that case, outsourcing is a good option. Utilities that are more advanced may want to learn from vendors that are cutting-edge in their analytics approach. Then too, cloud-based data management and analytics may be needed in order to meet the data management and computational demands of Big Data.
On the other hand, there are risks associated with outsourcing analytics. For example, the utility could lose in-house knowledge. Analytics initiatives could go awry if the service provider lacks understanding of the utility domain. Security and privacy of data is seen as an issue for cloud services. Transparency–the ability to view and understand the algorithms and underlying assumptions–is important. Regulators or investors may want transparency as well–think climate-related risk.
One utility analytics leader notes, “Utilities should be wary of the black box of analytics, both in terms of transparency and underlying data.” For example, a utility can provide separate data sets for asset management–PI data, work and asset management data–while the vendor can provide data readiness (quality checks, matching, consistency, etc.). If a utility chooses to change vendors, that data-readiness effort is lost and it will be difficult to replicate results using another platform.
Making the decision
When analytics are a high corporate priority, IT and analytics will be aligned. For example, one large investor-owned utility has developed a long-term model for strategic planning and governance. This utility decided to develop in-house analytics capabilities rather than outsource. Another utility is laser-focused on modernization and is looking at outsourcing low-value as well as high-value analytics. In some cases, analytics groups have more flexibility than does corporate IT.
Guarding against the risks
Just because risks are involved does not mean outsourcing should be off the table. Going outside may be the only option for utilities that lack the resources to do analytics in-house. Foresight helps guide against the risk.
In 2013, the Federal Reserve published a comprehensive guide on selecting and managing outsourced service providers. It covers topics such as due diligence, contract provisions, oversight, monitoring, business continuity and contingency plans. In addition to reading that resource, here are some other suggestions from utility analytics practitioners:
- Security and privacy. Scan data for personal and health data before it goes onto a cloud platform. Establish robust governance on sharing critical infrastructure data.
- Data transparency. Prepare the data in-house for data discovery before sending it to a platform. Include contract provisions to ensure transparency.
- Knowledge retention. Hire personnel skilled in analytics and analytics processes to actively manage outsourced initiatives. Make sure that in-house power, gas and water engineers are involved in applicable projects.
