IoT for Water Networks

[Title] Seizing the IoT opportunity. Telstra Vantage^TM  2019

PAUL FRANCIS (Head of IoT Business and Strategic Growth, Telstra): Yesterday, we talked about bridging the gap between hype and reality.

My colleague Jason Frankel and I talked about how if we don’t change our approach in some respects or if we don’t modify some of the approach that we’re taking, we could find ourselves more on islands of things rather than Internet of Things.

And we shared some mindsets and some frameworks and some ways of going about that, including the importance of things like interoperability and data sharing and standards.

Some of the things that laid the foundations in order to be able to scale, whereby you get not only the full value, the power of IoT, but you start to get some of the unexpected benefits of IoT once you integrate.

And once things move beyond their initial scope, we also talked about some of the proof points that are emerging and that we’re seeing being done at scale.

You can see many of them here.

So it gives me really great pleasure to introduce a couple of leaders in the field who will talk about some of their learnings, their journey, and share with you some things that hopefully you can take back to your own businesses as you undertake your own IoT adventures.

So please join me in welcoming to this stage, Sean Atchison, the CIO of SCT Logistics, and Mike Wassell from Sydney Water.

Welcome.

I could see you so why don’t we start, and Sean I’ll start with you.

Some people might not know too much about SCT Logistics, although we have a wonderful presentation here at Vantage, of course.

So why don’t you just tell us a little bit about the business to start off.

SEAN ARCHESON (Chief Information Officer, SCT Logistics): Okay. So SCT is a privately owned logistics and freight company.

We primarily move customers’ freight, which is contracted around Australia via the railways.

We own our own rolling stock, with 55 locos, 700 rail wagons, 500 containers, which are ours.

So it’s all massive amounts, on average, we move about 40,000 tons of freight every way you can, that will be refrigerated as well.

So, everything that you consume from a supermarket or from a warehouse-type environment, then we’re probably touching it.

PAUL: Yeah, fantastic.

A significant scale operation.

And Mike, I think most people would at least know the name Sydney Water, but perhaps they don’t appreciate some of the complexities in the scale that you’re dealing with.

Can you sort of paint a bit of a picture for us?

MIKE WASSEL (Head of Operational Technology Services, Sydney Water): I’ll give you a little bit of the picture.

Sydney Water supplies water and wastewater services to the Greater Sydney population, including the Blue Mountains and the Illawarra.

About $20 billion worth of assets; a thousand pumping stations, 250 reservoirs, 25,000km of wastewater pipe, 21,000km of water network, and about 4.5 to five million customers.

PAUL: Yeah, fantastic.

Now, Mike, just while we’re with you, I mean, I might be saying you’ve been in this game for a while.

What was it that’s sort of piqued your interest and started the journey around emerging technologies like IoT to help further address some of the challenges that you’ve been seeing, some of the complexities in the business?

MIKE: My background is I’ve done similar sort of technologies for a long time.

So we’ve worked in operational technology for a long time but we’ve looked after the trunk system, and that’s moving water around the network and wastewater around the network.

But what we don’t do is the distribution system.

So the 21,000km of pipe work, pretty much we rely on customers to tell us when something goes wrong.

And the same with the wastewater network.

So the first time that we were going to know about that is when a customer rings in and tells us something has gone pear-shaped.

The opportunity for us was really to be more customer-centric and to get ahead of the event.

Technology before I came along was really just too expensive to put sensors in that type of network.

So it was an opportunity of battery-powered technology, lower-priced LP ones that give us the ability to actually move it out into the network.

PAUL: Thank you.

A similar story for you, Sean, and I love a couple of the stories that you told me about the use cases or any examples that you saw, that made you think, or your managers think, “oh, okay, hang on a minute, we probably won’t get better visibility into our assets.”

Do you share that?

SEAN: Yeah, absolutely.

So as I said, we have a number of containers and one of our customers rang up and actually asked the commercial director if we were ever going to pick up one of our containers and it had been there for longer than 12 months.

So we’re kind of lost it in our own network.

And so he went, “okay, I’ll get it.”

And the next one, our fleet manager was driving past an auction house and saw one of our trailers and containers, just about to go to auction.

So that’s a real big kind of “whoops”, yeah, we’ve missed it, we’ve mislaid it.

We didn’t know where it was.

So when we looked at IoT, it became a no-brainer.

PAUL: And how have you been able to, in a new industry, which has embraced technology to a certain extent but also is quite traditional, in fact, both industries.

It’s sort of quite traditional in some regards.

How have you been able to illuminate the benefits to your management teams to be able to invest in this sort of stuff?

SEAN:  So we started our journey with Telstra and the IoT team December last year and we took a really small proof of concept, just ten devices, and that was five containers and the rest were trailers.

And within the three months, we actually saw the distances our containers were travelling around Australia and pretty much within three months, these five containers touched every one of our inland ports and had traveled sixty, 70,000km.

So it was quite a unique example for us.

And with the trailers, even though they’re only in Melbourne, again they travelled tens of thousands of kilometres around Melbourne.

So presenting that sort of data to the executive team highlighted why we needed to keep track of our assets and the opportunities that would arise.

PAUL: Yeah, they’re very powerful.

And Mike, sort of similar, but slightly different question in a way, but you know, the water area you built for long lifetimes, as in you have very challenging environments around gases and other things as well, with IoT and all of this being sort of emerging, how do you sort of build that into some of the thinking, how did you change some of the thinking about the way you approach this?

MIKE: Difficult question, Paul.

So it’s a work in progress.

The challenges that we’ve had around the environment are by far the most significant.

So IoT, when we first started with it, it was about digital metering and it very quickly turned into a lot of different use cases, so similar to Sean, we picked four use cases that were customer-centric.

Most of our challenges being around finding devices and sensors that will work in an environment that is challenging.

We’ve had between 20 and 80% failure rates.

So it’s been a journey of working out what works, what doesn’t work.

A lot of our sensors are below steel manholes, in a sewer with gases and other unsavory things that happen to be in sewers.

So it’s been a case of trying to weed out what worked, what didn’t work.

And the executive behind is pretty much all of the way, I think they see the value of it and demonstrating the benefits from a customer perspective.

It’s about moving from an asset-centric organisation, really, to a customer-centric organisation.

PAUL: Yeah, it gives you that visibility, so to take it back, one of the themes from yesterday was to be able to move to scale, you have to lay some of the foundations.

Obviously, you need to know what you’re aiming for and what good looks like in terms of the pilots and the trials that you’re doing, but you need to set up some conditions then that once you scale, whether that’s platforms or interoperability or other things.

So maybe Sean, can you sort of talk about that because you’ve now started to integrate, as we again spoke about yesterday, integrate some of the point solutions into more of a horizontal view, which is some of the aspirations, I think, if we want to get to with IoT.

SEAN: Yeah, absolutely.

We started off by just wanting to actually track the assets and what we’ve found is that our customers are actually interested in knowing where their freight actually is, not just the fact that it’s reached one of our terminals or we arrive to pick it up from their sites.

They actually want to track that asset, whether it be per day.

Even some of our customers are going down to the hour, especially on perishable goods.

So, what we’re now finding is that we’ve got lots of information coming through, how do we then bring it together and allow our customers to view it.

We have a number of customers, some of the largest in Australia, and they’re very concerned, especially perishables as I’ve said, where is it, what’s going on with it, and making sure that we deliver on time.

PAUL: And Mike, for you, in terms of how you thought about, as you touched on at the start, it’s a big, complex business.

There’s lots of different moving parts, lots of legacy in there, as well.

So how do you approach this with the thought of getting to scale, because at some point you sort of need to, I think.

MIKE: I think we sort of knew the scale when we started.

So, you have five million customers, you’ve got a lot of meters you’ve got to put up, 21,000km of… and we still don’t know the answers necessarily to all of the questions.

21,000km of water mains, 24,000km of wastewater… a lot of the challenges is around getting the right mix of sensors and technology to tell us what we don’t know at this stage.

So we don’t know whether you put one every kilometre, every 500 metres, every five kilometres, for instance, to get the right outcome.

We’re using a mix of IoT and machine learning to try… and artificial intelligence to try and work that out.

But from a scalability point of view, one of the reasons we chose the Telstra platform was because it’s scalable.

So we started from something that we knew would be scalable.

PAUL: So you laid a foundation on the platform…

MIKE: We laid a lot of foundation.

We tried 17 different vendors, five different solutions, five different wide area networks until we landed on the one that we were comfortable with.

PAUL: So let me follow that straight for a bit, because I know that you have sort of experimented with everything, both from the market and even building some of your own stuff as well.

What was it or how did you get to the point where you thought you had the evidence you needed to get to say, “okay, this is becoming clear for us which way we’re going to go.”

Choice is a real issue, I think, in the IoT ecosystem at the moment.

MILE: A big part of it was doing by learning.

So the reason we trialled all those technologies is there’s a lot of snake-oil salesmen around.

There’s a lot of people offering solutions to problems we didn’t know we had.

So we tested a lot of them.

We built some to determine whether or not the price point was right.

Or could we move the price point in the marketplace.

And the reason we finalised on when we’ve done enough trialling to be comfortable with the technology is that it actually done best out of it.

PAUL: Sean, Mike mentioned something interesting, which is the sort of the unexpected benefits of IoT in a way you don’t know what you don’t know, and when you surface the data and you start to sort of integrate it, you get some things which might be sort of aha moments.

What have you seen along the way?

SEAN: For us, it’s actually been very enlightening for what we’ve actually discovered and it’s really along the lines of how can we provide a more open-book service for our customers, actually showing that information to our customers, which then says actually we can now see what’s going on within our whole logistics supply chain.

So for us, it’s actually been really important and that’s a real side effect from what we’re doing.

We just started off by wanting to see where our assets were and now our customers are saying, “hey, there we go.”

But it’s knowing that we’ve got this huge amount of data and how, again, do we scale that up is huge for us.

And so that’s kind of the scary fact of it.

And it’s probably going to hit us harder as we get more and more information back from the IoT devices.

PAUL: Yeah, for some, I think it’s a little bit chicken-and-egg in terms of, you sort of feel like you want to know what you want to measure and what kind of outcomes.

But in some sense, you want to measure stuff and then you might be able to get some analysis that kind of gives you that.

And in most organisations, there’s no end of data anyway, multiple different systems, so have you kind of squared that because you say you’re going to be both.

You’re going to generating enormous amounts of additional data on top of all the stuff that you’re probably or possibly not looking at anyway.

SEAN: Yeah, so we’ve actually did a fact packet, just some IoT devices and we got around 10 terabytes a year because it’s pinging off every 120 seconds.

So it’s really scary when you think about that volume.

And then you merge all of your ERP data, all of your customer data together, and it just grows.

And how do you actually use it?

And in the logistics, that data is only really viable for however long the supply chain lasts, so for five to seven days.

And that’s a huge volume.

So what should we do with it afterwards and how do we analyse it?

And I think that’s where our trends are now going to come into it and the analysis of that data.

So we will then see other impacts of having that huge volume.

PAUL: So you’re starting to look back into patterns to allow you to then project…

SEAN: Exactly.

PAUL: I might make a sort of similar question for you.

I mean, you have obviously enormous amount of data coming out of the process control system and  others and you’re overlaying customer information, so how have you sort of dealt with the tsunami of additional data?

MIKE: We haven’t got to the stage where it’s a tsunami yet, but it’s growing.

The beauty of cloud storage is, you know, you can scale it up and down.

So, similar to SEAN: , we use the data when it’s valid.

So a lot of the challenges we’ve had is we didn’t have empirical data about what was actually happening in the system.

We had models.

So in terms of the data that we’re using, we’ll aggregate it for certain instances and we’ll dive deeper into it for others, and then we’ll throw it once we’ve got the learning.

PAUL: Yean, now I saw you present at, I think, it was Committee for Sydney event in Western Sydney a year ago.

And at that point, you were somewhat skeptical about the benefits of machine learning and AI at that point, in terms of being able to unlock some of those.

Are you still a skeptic or are you starting to see a little bit of…

MIKE: I don’t think I’ve ever been skeptical about it.

I think the challenge with predictive analytics was there’s an awful lot of companies as I said that will knock on your door and tell you they’ve got an answer for a solution that you didn’t know you needed.

But we’re getting good wins out of it right now.

So we’ve got a range of things and we’ve started employing people who are specialists in it.

And the real value is you need subject matter expertise.

So predictive analytics or analytics or machine learning on its own is interesting, but it’s only as useful as the subject matter experts that you put into it.

PAUL: Yeah, and are you finding it  a challenge to find people with the right skills to be able to…?

MIKE: It’s a challenge to find data scientists as opposed to, you know, a data analyst or at least… we’ve got a data scientist.

But it’s becoming more and more prevalent, so the unis have started to pick up… and it’s a journey.

PAUL: Yeah, big draw on the market for data scientists at the moment.

We want to come back to something you mentioned, SEAN: , because my understanding is you can have a scenario where one of your customers is saying I’d like to get visibility into where are my goods and where are they in the supply chain.

Well, of course you’re carrying stuff from different customers.

So how are you dealing with sort of just slicing the data in a way and getting the sharing that they’re asking for without giving away data that you don’t… which was your… perhaps you can’t share from other customers.

SEAN: Yeah, so it’s a really good point because you want to service all your customers in the same way.

But we don’t always do full loads and so we have to be very careful on how we do it.

So we’ve actually partnered with Telstra Innovation Labs to actually provide us with some of the analytical data and how we merge it together.

So that’s just starting for us and what we need to make sure we do is we’ve got data integrity and from that how do we then divide up what’s for customer A and for customer B, especially where we then share rail wagons or containers and we merge the freight together.

And that’s gonna be the really hard one because we don’t want to expose the wrong data to a customer and that’s going to be our biggest hurdle to move forward with.

PAUL: Yeah, so having solved that brokerage, through broker relationship almost in the middle might help to alleviate some of that.

And Mike, I think you’re sharing or starting to share some information with local councils, perhaps in the future with our metropolis in Western Sydney and others.

Similar sort of question, how do you deal with things like personal identifiable information or information that becomes identifiable once you merge together different sources of information?

MIKE: Yeah, we’re right at the beginning of our data sharing vertical.

So it’s a 50-50 split within the organisation about whether we give the data or share the data with anybody or we give access to it privately.

So we’ve got a desire to share with councils and take data from councils.

We’ve just started that journey.

But we certainly anonymise the data regardless of whatever we share.

So there’s no personal data given out at this stage.

PAUL: And the view for those who don’t want to share is because they see value or just because the risk factor in doing so?

MIKE: It’s critical infrastructure, so it’s a risk aspect.

So the more information, especially when it’s geo-spatially located, it’s a challenge, you know, with some of the data that you give up and all of the solutions that we’ve put in are aimed at a customer or an environmental outcome.

So the last thing we want to do is lose customer sentiment over giving the data away without their knowledge.

PAUL: Yeah, makes sense.

Let’s talk about that for a second because some of the higher order benefits of IoT, aside from the operational efficiency in the asset management and so on come from being able to improve customer experience whether that’s visibility or that’s proactively informing people or knowing when a leak in which properties might be affected.

So how are you approaching that with this sort of customer experience lens as much as the operational efficiency side?

MIKE: I think from when we started IoT, it’s entire ethos was to be about the customer.

So at the same time, IoT is just a technology that we’re using as part of a major customer transformation within the business.

We started a thing called Customer Hub, so Customer Hub takes our 24X7 operations centre, our call centre for any calls that come in, any calls about faults, and dispatchers for planning and distributing work, and IoT is just an enabler into that technology.

So it’s really about the customer journey as opposed to IoT for the sake of IoT.

PAUL: Similar question, SEAN: , in terms of the benefits that you’re in charge, but you touched on it in terms of the transparency you’re providing.

How do you sort of approach in the customer experience lens?

SEAN: For us, everything we do is customer-focused, and providing them with real-time access to where their freight is is really important.

And it’s different from us as consumers, where we would see Australia Post just right here, right there.

It’s actually about real-time, so they can have the ability to redirect freight if they need to.

So they have actually asked us to offload freight at one of our depots as the train goes through, so they can service other parts of their business.

So lots of opportunities will arise where our customers know exactly where their freight is, because it’s not ours.

So that’s kind of where we’re going to be going.

PAUL: So it’s interesting… so it’s just a two-way thing they can help sort of tune your system in a way by providing you or request alternatives, perhaps might be a better way to put it.

SEAN: Yeah, so we actually have customers that will say, “can you offload one, five, and nine pallets and then take the other pallet somewhere else.

And so that will give them the capability and the ability to almost dictate on the fly where their freight will go.

So for us, there’s a seamless… it’s endless for us what we could do with being able to monitor it.

PAUL: And that’s some of the integration that you would then have with your enterprise systems around scheduling or optimisation and so on as well.

SEAN: Yeah, so we will then look to combining the freight data with our train services later and that’s another area that we’re working with Telstra on and especially around bringing everything together.

PAUL: As we all know, it’s not sort of straight line in terms of the journey.

There’s plenty of hurdles along the way.

Mike, what are some of the key things that maybe if you had your time again and what you know you’d deliver differently, but maybe some lessons for the folks here in terms of some of the key gotchas that you saw and things that helped you kind of get over those hurdles?

MIKE: Yeah, so key gotchas were the environment that we’ve put the equipment in, I might have tested things in the field before we done them in the labs, if I’m doing it again.

So we tested everything quite rigorously in a laboratory but when we put it in the field, we found a lot of things didn’t work quite the way we expected to do when it was in the field environment.

Other than that, I’m quite happy with the way we approached it.

You’ll learn just as much from your failures as you do from your successes, so I don’t think I’ve got any regrets in that respect.

PAUL: That’s good.

Sean , what have you kind of learned along the way?

SEAN: For me, it was very much about the journey I took the board on, and we started small and I never expected the response to be just, go and do it.

I expected more of “okay, let’s roll it out slowly and all of a sudden, why wasn’t it all done?

So the expectation of what we could do were all resolved and it’s not all about taking national business on that journey.

The board said “okay, just go and do it, Sean.”

And I’m like, okay.

And then it was sudden, that’s happened, and then all of a sudden you’ve got huge amounts of data and where we wanted to be next year, we’re now here.

So it’s really about taking it slowly, steadily, and building up rather than coming into one.

PAUL: Let’s dive a little bit deeper into that both of you had to sort of top-down imprimatur around difgital or customer experience and see for the board.

But I think bottom-up and top-down sort of works best in some regards because you can’t just… what experience have you had with those who are out in the field, whether they’re the drivers or the engineers or others?

Has there been any resistance in terms of automation or machines are taking over, Big Brother is watching me, or any of the things that you might have thought you saw?

SEAN: We’ve had a lot of feedback.

This is only one of the Telstra products that we’ve actually deployed in the field and with some of the other ones, the business has come back and said, we actually, we’d like it, our drivers want more of it, but it’s the way you actually sell it.

So with some of the products that we’ve deployed, it’s very much about safety and looking after the health and wellbeing of our drivers and our train crews.

And then it’s how do they want to use it.

We have a pre-conceived idea and then it’s about their feedback and then kind of making it better for them.

PAUL: Right, so they have the opportunity to actually help to shape absolutely the way decisions are used.

And MIKE: , sort of similar question for you in terms of the engineers or field services or others.

Has there been any pushback from your side?

MIKE: Not really.

We’ve had quite a lot of support.

So the company had already had a fair transformation with SCADA before that.

So they were used to the fact that everything was automated.

The system works automatically now.

Most of the things that we’re deploying with IoT is about making our staff or our customers’ jobs easier or better.

So the staff are quite supportive of it.

It stops them having to go into environments that they would otherwise have to do fairly unpleasant, not the nicest things to do.

And typically you go in the middle of the night when nothing’s happening.

If we can stick a sensor down there and get the same outcome, that’s a win-win for everybody.

PAUL: Yeah.

And in terms of sort of innovation within the broader business and how you’ve kind of fostered a similar passion that you’ve both got for this and bringing people along for the journey is one way to say it.

How do you kind of engage change management processes that you’ve done the stakeholder engagement types of process, that they get real visibility into what it is you’re trying to achieve?

SEAN: We’ve engaged at both ends of the spectrum, so we engaged the board for their support and we kind of missed out in the middle layer and we go to the people whose daily jobs and they have to interact with the technology that we’re rolling out and we get them on board.

Once they’re on board, and in this kind of age, digital age, we’re actually engaged with the older generation.

So when we rolled out some technology in our vehicles, we actually take the drivers who are in their 50s and 60s who might still have the old button phone and get them involved.

And then once they do it, then the message gets passed really quickly.

It’s not big and scary, it’s actually easy to use and interactive.

So we work both ends.

PAUL: So you have some champions and some unlikely champions, perhaps.

SEAN: Absolutely.

PAUL: Mike, for yourself, how…

MIKE: Similar story.

So the board were on board from the beginning, but we’ve done stakeholder engagement across the business.

And it was more about ideation, so we started with the field crews and the wider stakeholder group.

We had 112 use cases that came up and we prioritised those basically against the organisational goals and plans that we had, which was all about if it was customer, environment safety, it got a big tick.

And then about the complexity of actually implementing it.

So for the change management perspective, most of the people were on board.

The biggest change, surprisingly, that we’ve had to struggle with is digital metering.

So that’s a journey that we’re still going through.

PAUL: And that’s from the consumer perspective or from internal, within the organisation?

MIKE: Both.

So customers, when we went out and done customer surveys.

It depends how you put the question to them.

So they are not overly keen on looking at the water meter every day.

Now it’s not the ideal use but one of our significant changes if we can change customer user patterns by five percent or less.

That’s five percent of infrastructure we don’t have to build in Sydney is growing at a significantly fast rate.

PAUL: Yes, and current water restrictions and other things about drought doesn’t help here.

Now, I just want to finish on, I know we sort of shared some lessons but if you could share one thing to the folks here to take away or that one thing be…

SEAN: Sure.

Start small and build the momentum.

PAUL: Excellent. And Mike?

MIKE: Mine would be, be clear on the business problem that you’re trying to solve before you start.

So I see a lot of people put in IoT because it’s the buzzword just now.

But if you understand the problem trying to get to a good place to start.

PAUL: Fantastic.

So hopefully you’ve seen that whether you’re in aviation or advanced manufacturing or retail or residential that there are some lessons that you can take back into your own businesses here.

We’d actually love to find out more about what you’re up to in some of the challenges you’re facing.

So please continue to talk to us but it just remains for me to thank my wonderful panelists so please join me.

Give it up for Mike and Sean.

[Title] Telstra Vantage^TM. Insight. Inspiration. Innovation.