How are asset-heavy industries monetising data? Lessons for utilities

Data analytics is transforming other asset-based industries' daily operations and commercial models. How could it impact the utility sector?
Published: Thu 07 Dec 2017

Utilities are relative newcomers to commercial competition having been cushioned by state subsidies and market monopoly.

Not so their asset-intensive peers. Manufacturing, automotive and aerospace have long operated in a climate of intense competition, margin protection and operational efficiency.

Unsurprising then that these industries have a sharper, more granular view of how they operate and a more mature approach to data.

There is a clear progression from the collection and analysis of data for specific use cases and wider, transformational business model change.

Take the automotive sector, says Iain Stewart, Practice Partner, Utilities & Smart Cities, at Teradata, a business analytics and data platform company that works across multiple industries.

Stewart says: "Data analytics is driving efficiencies in manufacturing and servicing in the automotive sector, while also enabling a shift in the way we as individuals buy from and interact with dealers. They are moving towards selling leasing and service-based contracts versus just selling us cars."

He adds: “By collecting and analysing data from multiple sources, automotive companies are able to work out what drives costs in their organisations and how to maximise future revenues.” 

Broad-ranging analytics improve efficiency and reliability in areas as diverse as supply chain optimisation to predicting secondary faults likely to occur soon after an initial problem - thereby reducing repeat servicing visits.

Over time, those same analytics capabilities will underpin a wider transition into smart mobility - the ultimate mixed transport model for the future that puts the customer squarely at the centre.

This is all only possible if you approach data analysis in the right way, says Stewart.

“The key thing that utilities can learn from other asset-intensive industries is to invest in specific use cases that will drive value quickly and demonstrate the benefit of analytics.” 

“At the same time," he advises, "they should take a strategic approach to managing data that can create an enterprise-wide data-driven strategy but focus first on just the data to address specific pain points and generate an initial return on investment.”

Wider manufacturing business model change

Stewart points to general manufacturing as another example of using specific analytics use cases to identify costs within a business and drive operational transformation.

By shining a light on the detailed costs of producing goods and delivering ancillary services, the manufacturing sector is also moving towards more holistic service-based contract models.

There are higher margins to be made in providing contracts with bundled services or even entirely made up of services, without any asset capital sale than building and selling assets alone and hoping for ongoing support contracts.  

But only if a vendor understands and can reliably forecast the future cost of delivering on these services and contracts, says Stewart.

He again highlights this as a key learning for utilities - it is hard to transform your model successfully if you don’t understand your costs.

“The solution is to take data seriously, and over time integrate data on a single enterprise analytics platform. Deliver those quick wins. And as further data is integrated, deliver new insights again and again - perhaps even in areas you had never expected.”

In this video interview, Teradata's Iain Stewart reflects on how the utility sector's handling and perception of data have shifted in the past few years. Stewart also explains how data science is playing a role in predictive maintenance for utility assets. 

Integrated data provides a way to understand not only an individual asset’s behaviour and associated costs, but also grouping of similar assets allowing for peer group benchmarks.

"You can quickly analyse other assets that may experience the same behaviour," explains Stewart.

"For example, based on additional data about operating conditions, assets can be classified into groups to track any degradation- such as 'critical - must replace immediately', 'next time repair', 'replace at next scheduled inspection' or 'observe - keep on monitoring'.

Predictive maintenances - lessons from aerospace

Mission critical data is one such example where analytics can move from condition-based to predictive and even prescriptive maintenance regimes.

Take the aerospace industry, says Stewart, where safety and operational efficiency are paramount. Just as a utility has to keep the network online with minimal interruption, so does an airline to avoid service disruption.

To manage safety and to combat revenue losses through asset failure, the aerospace industry has deployed sensors on aircraft critical infrastructure.

But data generated from thousands of sensors, generating readings as frequently as every second, only delivers value when it is analysed to find indicators of future performance degradation; inefficiencies; differences across fleets etc.

This is a high-stakes game. Boeing and Carnegie Mellon University launched the Aerospace Data Analytics Lab in a bid to use machine learning to detect value from masses of operational data. Teradata recently partnered with GE Aerospace to create a comprehensive analytics environment to service all aspects of the aerospace industry.

Stewart says: “Transferring these capabilities to a utility scenario, instead of a utility reacting to outages as they always have, a data-driven utility will know when a network or asset is going to fail because their analytics have spotted an anomaly.  

"At the very least, that may have triggered an alarm in plenty of time for an engineer to take actions that keep customers connected. But the opportunity is there for much more - even for proactive, machine-driven interventions.”

Data about your assets provide means for working on predictive analytics. Predicting when a component is likely to fail is only one element in a broader strategy underpinned by analytics.

"Combining the asset information with your supply chain and logistics data generates suitable time windows for performing the repair or replacement.

"But it doesn’t stop there," says Stewart. "In a service model where you sell uptime, it is important to make sure the replacement is successful.

"By combining the asset data, supply and logistics data with HR training data about service technicians, you can ensure that you have the right person with a track record of successful repairs doing the work.

It is just this shift from reactive to predictive that other industries have made and are still making - both in terms of safeguarding revenue and better managing assets. This might be the greatest lesson utilities can learn.

Stewart concludes: “Utilities will argue that they have always used data, and of course they have. But they have never used it in this way before. The commercial need is now there to cut costs and offer services to drive new revenues. The solution to these challenges lies in data and analytics.”