Panel Discussion: Using Data and Analytics to Prevent Maritime Incidents

Data and Analytics to Prevent Maritime Incidents– The Posidonia conference is a hub of innovation for the maritime industry. It is a perfect fit for HiLo and our mission to deliver the best life-saving technology, analysis, and collaboration.

We brought together a panel of 4 highly respected industry experts to talk about how to use data sharing to save lives, talking about:

• Different approaches, tips and challenges
• Helping mariners
• Finding the right data
• Standardisation
• Their visions of the future.

Panellists

• Mark Warner (MW), Director – Maritime Performance Services, Lloyd’s Register
• George Kouloukas (GK), Safety Manager at Gaslog LNG Services Ltd
• Andy Cross (AC), Chief Operating Officer at HiLo
• Stuart Edmonston (SE), UKP&I Loss Prevention Director

Q: What does data mean to you?

• Mark Warner (MW): To me, it’s factual information that can be collected, aggregated and analysed, and of course acted upon.
• George Kouloukas (GK): For us at GasLog, data is the bone marrow of our daily business.
• Andy Cross (AC): It’s any event, anything you can measure and collect to use to further your company.
• Stuart Edmonston (SE): Very briefly and very quickly: data is something to share and learn from.

 

How Do You Use Data?

Q: What else do you do with your data – how do you use it?

GK: Basically, at this moment, data drives the business decisions, we are using safety data for our campaigns.

So instead of working on the incidents that we might have, we prefer to see our data and prevent them. We run campaigns in advance, before we have an incident, from the hints we are taking from the data collection.

 

Q: How does the Class assurance work?

MW: To build the AI register, which we were the first Class society to launch, we’ve had a number of discussions both with ship owners/operators and providers of AI, including digital twins and digital health management.

And so we thought it would be good to establish an AI insurance framework around that, and also a certification level for those applications. We’ve now certified 7 different applications and we’re working with companies such as FURUNO who are creating digital health management for our digital twins.

So that will continue to establish and we’ll probably move that out into other areas, like optimisations, where AI is of use too.

 

Q: Stuart, how does UK P&I loss prevention handle the conflict between business and insurance?

SE: Well, I think my charts suggest that we’ll never be out of business because the cost of claims keeps going up and up and up.

But the relationship we have with a member is a totally different relationship to the underwriters’ relationship with a member of the club. Of course, we work together, but the essence of being a member of a big international Group club is what it says: it’s a club.

You want your claims to be paid by the insurance company. The loss prevention department is there as a sort of assist.

We have some members who need our help more than others. We insure pretty much everything that floats on the shipping side for Shell – I would suggest they don’t really need our help in safety.

But there are many members who do need our help. And what we can do is disseminate the learnings we have from the claims we have, and invite them to a course to help them improve on human performance, which is causing the claims.

So, there is no direct correlation between needing loss prevention or using loss prevention and increasing your premiums. It is very much the case that if you have lots of claims, it makes sense that your premium will go up, but that is analysed at the end of the year with the underwriters.

What we do is identify the member that has a problem with claims. Mostly these big claims that we deal with are entirely fortuitous – it can happen to anyone, it can happen to Shell, it can happen to the guy running a 20-year-old bulk carrier, it can happen to anyone. But generally, if it’s cargo claims, we work with that member to try and help them reduce the chances of those claims.

It’s very much a partnership. It’s a club, and it’s an asset. You join a good club, you’ll have a loss prevention service, and you’ll have a team of lawyers to help you with your claims.

We want to pay the claim, but we also want to mitigate the claim, or importantly prevent those claims happening again.

 

What is Causing Safety Issues?

Q: Where do you see the root causes: onboard or onshore?

GK: For sure, Greeks have an excellent saying for this – the fish always rots from the head! So you need either onshore or onboard to be data gathering and to set a culture of clear reporting. If you don’t have a culture of clear reporting either on the vessel or through the office, the only thing you’re going to have is meaningless data which you’re going to use in order to fill up business data. But at the end, it’s not the data you want.

Why Should We Share Data?

Q: How do you overcome the challenge of data sharing?

SE: I think it starts with these types of discussions. 5 years ago, even 3 years ago, you’d never have thought that all the International Group Clubs, 13 of them – that’s 90% of world shipping – no-one would have thought they would be starting to share safety data. That’s only happened in the last 2 or 3 years.

There’s a report on the International Group website about pilot incidents, that started with clubs bringing all their incidents into a database and discussing the cause. We’ve just finished a collection on mooring deck injuries, there’ll be a report out on that. Awful numbers, just short of 700 injuries on mooring decks in the last 5 years. In any other industry it would have been wiped out, you just wouldn’t be allowed, you wouldn’t have airport ground staff hurting themselves, regularly, like they do in the maritime industry.

I think these types of collaborations are slowly starting to pick up. As I say, the International group gets a lot of criticism for not working together, but I tell you something, there’s a big uptick in activity – sharing safety data to try and stop these things happening. Even if you’re cynical, thinking we’re only doing it to save dollars, it’s still saving lives. So, I think we’re getting there, it just takes more time.

The aviation industry is a great example. They’ve got a taxonomy for incident investigations, so there’s standardisation out there. It doesn’t matter where you are in the world, or what plane you operate – if there’s an incident, it fits into that word, or that column.

We need to get there, and we are getting there. I’m confident.

GK: Basically, the problem is that most company owners are reluctant to send data for one reason: they’re afraid of commercial implications this might have. So for sure, you need big trust with the one you are sharing the data with.

Of course, it’s important for an organisation to have a growth mindset. If you have a growth mindset then you learn from your mistakes. You share them with the rest of the industry. You’re not afraid of them. You hold onto those, you get stronger, and you continue.

What are the barriers to Data Sharing?

Q: Do you face challenges in terms of collaboration across so many organisations?

GK: Yes. For us it is a challenge because we need to feed all this data into different organisations. At first, we achieved the centralisation of our reports. We managed to centralise our reports in one place, we managed to have only one database to deal with. And now, for us, it’s easy, because it’s just an export and we can share it with everybody. 

Then, the biggest challenge we have faced is the information we gather from the old databases, in order to bring them into the data warehouse, and make them comply with the new structure of the data.

From now on we have a fixed policy regarding data:

• What kind of data we will need
• What we’re going to use
• Why we need them – because we keep information about something, and who is asking for this? We didn’t know, someone asked us 10 million years ago…

But now we have a policy on data, we have centralised data, it’s very easy

to feed everybody with the live data.

How Do You Find the Right Data?

Q: How can you be specific enough to understand the real problems? You said the majority of issues are navigational? What is causing that?

SE: I think around 89% of issues are navigation failures, whether it was a bad reaction to a problem on the ship, or whether it was just a communication issue with the pilot. The Ever Given is a great example of nothing going wrong with the vessel, but a massive claim and that litigation will go on for years and years to come.

That’s definitely an area for us to concentrate on. Thanks to the albeit very high-level data, we can see it’s the navigation errors that are causing the clubs money and causing the increase in premiums. So we’re concentrating on human performance, generally on the bridge.

 

Q: Do you think there is enough data in this industry when it comes to analytics?

GK: For sure there is a lot of data in the industry, and each company collects data.

The problem – and it’s something that I found with my organisation a couple of years ago – is the data does not have the correct structure to be analysed. This is the biggest problem.

So you need to streamline your data, put in another framework, and have the standardisation in order to be ready for analysing.

For sure, we can have millions of records, but we didn’t know what to do with them.

MW: I hate to say it but I think I had the data standardisation question on a panel 23 years ago!

In terms of availability of data, there’s still areas where no-one has a complete understanding of all data and where and how it should be used. Obviously ship owners are there and have good knowledge of the source.

But I’ll just give you an example: we’ve been developing software to look at powermetric rolling, which is still an issue. And I started talking to the team developing it and they said: ‘We’ve got to get the hardware on board, we’ve got to get it installed, we’ve got to get it manufactured, as that’s the only way we can measure the pitch and roll data’. I said: ‘Well, have you looked at other sources? Coming from Inmarsat, I know the antennae manufacturers have got it,’ and they said: ‘Have they?’

You bring them together, and suddenly you’re working in collaboration and in partnership. What other companies or other sources are there for data? I think we’re starting to break through there, but it’s going to take work, and of course that requires collaboration as well.

 

Q: Stuart, you mentioned that you don’t see all of the data from ships you insure. What are you doing in terms of creating that trust so you can protect vessels?

SE: It’s a bit of a difficult question. We do have trust with our members. It’s more that if for example, there’s a claim where we are bang to rights, the ship is 100% liable and admits it, the member wouldn’t thank us for spending us a lot of money on local lawyers and English lawyers to try and work out what actually happened and learn from that, because it’s all adding to their claim.

So I see that a lot of the incentive is on the member, the ship owner, to properly investigate – or the flag state. It could well be that we don’t as a club investigate an incident, but we do expect to be updated on root cause analysis, any follow-up, or activities on board.

But generally, if it’s very much to the benefit of the owner we do investigate it, we do throw money at it. A good example of that would be a collision, where liability is proportionate. That’s where we do learn a lot about the cause, because of the money spent on it by a large law firm who are trying their best to mitigate losses.

But if it was a $50,000 claim in a port in Manila, where a ship has hit a navigation buoy, it could have been a lot worse. It’s hit a navigation buoy, but it could have hit a rock. We wouldn’t investigate that as a club, we’d expect the member to investigate and be forthcoming with their findings. We’re not police, we’re insurance. We can’t be too policeman-like.

 

Q: Should we capture the subjective experience to supplement objective analytics? If yes, how can it be done?

GK: This is a matter of urgency and import. If you have open-end tools, you’re going to have stories, so you need to design a tool not to allow anyone to put subjectivity in the report – you need to be objective.

It’s very difficult. You cannot try to measure oranges if you count lemons. You need to spend time, effort and money in order for the tool to be objective and have a meaningful outcome. Otherwise, it’s opinion.

MW: I totally agree with what George is saying: you do have to have that objective, ideally, standardised data.

But I do think there’s also value in subjective experience. Going back to Andy’s phone analogy, I’m sure people here use Amazon. I’m sure people read the reviews. It gives them an idea in decisions: are you going to buy that 1-star versus that 5-star?

That subjectivity is trust-based. The company I work for are doing a lot around trust and trust ranking, and dealing with suppliers and even shipping companies as well.

It shouldn’t be ignored. It’s got to be captured. Going back to Andy’s slides, trust is a very important element, and you get that from subjectivity.

How Do We Standardise Data?

Q: What’s your view on standardisation of data?

AC: Standardisation of data – that’s another issue we have!

What we see is over 500 different terminologies for the same data set, so each company, each entity, puts a different name for the same data. That’s the first problem, you have to identify what the data is, and give it the same name.

That’s the easy part! The hard part is changing everyone to use that same name.

Only when it’s standardised can you do anything with it. Before it’s standardised it’s a complete mess. And that’s the situation shipping unfortunately finds itself in now. Over the years, as you try to grow, you try to make sense of it, but I guarantee there’s a lot of people making more of a mess of it than getting towards standardisation.

 

Q: How do you actually start standardisation?

AC: With great difficulty! As we’ve been saying, 23 years, 25 years…

There’s more and more technology that companies get in their own systems, and they all have different ERPs, with different capabilities. Some are very editable.

So the challenge is: even if there is a list of standard terminology that we should use to classify the data, a lot of these systems are open. If your HSEQ department or the superintendents get a report in from one of the ships, and it doesn’t fit into the boxes they have already, they’ll make another one up.

We keep creating these boxes of new terminology. We’ll never move forward unless we agree to stop with a certain bunch of terminology. There’ll always be a few things that will go into your ‘others’ category, which is a terrible category to try to do something with.

It’s a problem we can only solve together. This whole discussion has been about being together. Bringing data together. Bringing thought processes together. But there is only one way that I can see, one solution – and that’s as an industry and not as individual entities.

 

Q: Is there a standard tool to assess or grade risk?

AC: There is no one tool at the moment. I can confidently say that because no-one – and no-one on this panel – takes all of the data that we’ve been speaking about, and processes it and through a single. There’s no one standard.

It takes people a lot to change. It takes a few years within your own organisations to change the name of something, to change the name of a procedure.

There is nothing. And I think this is where we’re trying to get to, to make everyone in this room believe in it, and come with us on the journey that we’re all on here.

Is data replacing seafarers?

Q:  There is so much data available. How can we help shore staff and ship staff to find the right information?

AC: I think we go back to the decision support system. If you look at where we are now, certainly I think in 10-15 years I won’t be in the industry, many people will have retired.

The generation coming up behind us definitely will need a mobile phone in their pocket to make their decisions: life decisions; decisions on board; in their daily work. I think that’s where the organisations should head.

It’s not about the officers, the vessel managers. People are trained on board to make decisions. The officers don’t make every decision. It’s the seafarers, the management team that make the decisions, and they have to have that [decision-making] tool.

MW: Only true digital natives understand digitalisation and digital transformation. Probably across the board, not just at shipping companies, at the moment you probably have 20%. Hopefully we’ll be at 100% one day. Then you have the real capabilities, whether that’s AI, or whether that’s reality. I see ships going like cars, where when you’re coming into port you’ve got everything around you, in augmented reality.

But you’ve got to have digital talent across the organisation. Until you have that we’re not going to see the full benefits of digitalisation.

SE: What we need to make sure is that we don’t leave the seafarers behind. We’ve got human beings on ships. What we can’t do is to talk about digitalisation and collecting data and taxonomies and whatever else, and leave the seafarers with a go-to guide on the ship that’s sent out to them in the post – ‘this is what you have to do’.

They need to be part of the business. I think we’ve come on since I was at sea:

I never really felt part of the business at sea. There are companies now that are doing really well in so far as their captains and the officers for sure are really part of the business. They know why they need to get to this port at a certain time. They understand the needs of the charterers because it’s all coming down from the CEO or whoever, and they’re a very important part of the business of shipping. 

I think that’s where it won’t happen in the next 5 years, if we’re not careful. We’ll have these wonderful results and collaboration shoreside, but we’ll have the really important bit, the people on the ships not really understanding – what the hell does data mean? What is it?

I’m a bit concerned that, like a lot of things in shipping, we forget that the ships that pretty much drive this industry have human beings on board. And they don’t necessarily – no fault of their own – understand what we’re talking about here. I think in the next 5 years we need to be careful not to leave them behind.

AC: I completely agree with what Stuart said. The seafarer is at the heart of it all.

What Does the Future look like?

Q: In the next 5, 10, 20 years, is there a crazy idea in your head where you see the shipping industry going – something new or different; any predictions?

GK: For sure we will see digitalisation driving the shipping industry. I’m 100% convinced that in the future companies need to digitalise their operations on board and on shore in order to be efficient. This is something that is inevitable.

AC: If we’re talking crazy ideas, I think in the next 5 years I think that’s possible.

There are 4 halls [at the Posidonia conference] of traditional ideas and some crazy technology out there. I’m sure a few of these booths will come together and bring some of this technology together, and that’ll be the start of the future.

You can’t be looking through a VR headset at the aft end of the ship when there’s a fire at the forward end of the ship. It all has to come together.

 

Q: To Mark, what are the goals of the Class in respecting data and protecting its safety? Do you have any vision of the future in this space?  

MW: Yes, so obviously that was one of the reasons why we commissioned the report in terms of defining the space when it comes to artificial intelligence and machine learning.

I think the 3 areas that I covered – the new technologies, AI and machine learning – but obviously you’ve got big data analytics, wearable technology, even drone use and the data you can gain from that, it’s all things that we’re looking at.

In terms of the classification society, we have an innovation division, and our performance services is based on digital data as well. So they’re all areas we are looking at.

But I don’t think you can underestimate that need for collaboration, that sharing data. Also, when you have that human factor, you need to pay attention to how you implement digital solutions, in the best way, to ensure every area is covered and that you’re not introducing new processes and workflows to people who don’t understand.