Skip to 0 minutes and 4 secondsHi, I'm Elizabeth Hollinger and I'm head of analytics and BI for Aggreko. So my team is comprised of three different parts. First of all, our data engineers who ingest information into our data lake that we can use to build operational reports or predictive analytical models. The second part of the team are our data scientists who use that data and build predictive models for use across our organization. And the last part of the team are the BI team and that's a mixture of developers and analysts who produce operational reports that are used all across the globe.
Skip to 0 minutes and 33 secondsSo Aggreko started their analytics journey around two years ago and it was about the same time that we started to connect our fleet of assets. So we have around about 100,000 generators that operate all around the globe and we fitted them with IoT sensors. So what we did a couple of years ago, we started to stream that data in real time and analyze it using analytical techniques to understand whether a particular asset and a particular region was operating anomalously. And we did this for a couple of reasons. First of all to improve and make sure health and safety was really at the forefront of what we're doing.
Skip to 1 minute and 8 secondsBut second of all, to make sure that we didn't have any instances of unplanned downtime for any of our customers. So we started to build some of these anomaly detection models to pre-empt when a machine might shut down, when it might go on fire, when a particular part of the asset isn't performing the way we want to and we have to swap it out. And we kind of established those models 18 months or so ago and they've been running for a long period of time. And then around about a year ago, when those things were established in the organization and really starting to drive benefit, we started to broaden our horizons and what we were thinking about in analytics.
Skip to 1 minute and 46 secondsSo we'd already thought about optimizing that performance of the asset and we switched into thinking, well, what can we do about identifying our most profitable customers? How can we think about how we optimize our logistics process? And so over the course of the last six months to a year, we've really started to extend that analytics remit and support all of the domains across our global organization. So where we see data, AI, analytics and automation really playing a part for Aggreko is in enhancing decisions that human beings are already making.
Skip to 2 minutes and 17 secondsWe don't really view that as replacing a human being or replacing the decision that they might make, but actually we see using data analytics, using BI, using insight as a way of aggregating and summarizing everything we know that's happened about an asset, about our sales, about our customers in the past and to summarize a version of that information that a decision-maker can take, they can look at and analyze and make a better and augmented decision by using that. So in Aggreko we don't really see generally data replacing human beings, but we just see it as enhancing and augmenting the decision-making.
Skip to 2 minutes and 51 secondsAs I mentioned, Aggreko started our analytics journey a couple of years ago and we built our very first anomaly detection model and that was fantastic. So we were sending out alerts in real time to technicians in the field to say, "Have a look at this asset. We think it's not behaving as it should be." And it was fantastic. But as we started to add more and more modules onto that, we were sending more and more alerts out to those technicians. And the benefit of having those alerts in real time is that we can take quick and preventative action when we need to.
Skip to 3 minutes and 18 secondsBut the downside is that we need to prioritize the things we're going to look at first and make sure that we don't overload with information, but prioritize the information that is most helpful for that technician for that point in time. So I think there are definitely pros and cons and we've been working really closely with all of our engineers, in collaboration with our technicians to make sure that the information and insight that we provide to them is most helpful in enabling them to make the best decision.
Skip to 3 minutes and 47 secondsSo I think it's always difficult when you try something new in an organization when people are used to making decisions in a particular way and they don't want to move away from that status quo and they don't see the benefit really of doing that. So what is really pivotal for us in Aggreko is that any of the projects we pick up, we lead in collaboration with one of our business SMEs. So it's really important that they define the question that they want to be able to answer from the data.
Skip to 4 minutes and 10 secondsAnd we work with them collaboratively and iteratively to be able to understand what the data is telling us about that question and then how they can use that to enhance the decisions that they are trying to make. And it was difficult at first. We did get some pushback and people said, "But we know this already. Well, you can't tell us any more than we already understand." But one of the pivotal moments for us came when we predicted that one of our particular assets was going to go on fire. And we said, "Based on all of this history and the way it's been operating over the last couple of days, we think there's a really big risk that this will happen.
Skip to 4 minutes and 46 secondsPrioritize someone to go and look at it." And when they got there and shut that asset down and opened it up, they found that one of the pipes was smoking at that point in time. So that for us was a really key moment to say, "Well, actually we're enhancing what you're doing already, we're flagging the right things, and we're really helping you to make the best decisions that you can."
Aggreko case study
Aggreko are suppliers of power and temperature solutions globally. I spoke to Elizabeth Hollinger, Head of Analytics and BI at Aggreko, and asked her if he could give us an overview of the Aggreko’s data journey and experiences. I also asked her about where they drive value and how employees are being taken on the journey.
One of the many great things they have achieved through the better use data was the work they did to change how they meet customers’ needs, they moved from being viewed as a supplier of generators to a supplier of an unbroken power supply.
The data insights Aggreko generated gave them the confidence to make change their offering, they effectively moved from being a commodity provider to become a partner.
I hope you found that as interesting as I did. I particularly liked Elisabeth’s views on engaging employees across the organisation through conversation and collaboration. I also found the insights on balancing managing real-time alerts with employee overload fascinating.
Elisabeth and the team have achieved great things by changing the way Aggreko help customers and in doing so changing the market dynamics. I initially saw this as an improving operations case study, but these changes were only possible after everyone trusted the data enough to make business critical decisions using it.
Aggreko decision and operational improvements. (Click to expand)
We will hear more from Elizabeth in week 3 when we explore the wider aspects of data maturity.
Have a look in the “See Also” section below for more information on this project.
© The Data Lab & Seek And See Ltd