Fleet Performance in Shipping - a high level overview of how to evaluate vendors

For the past 6+ years, I have been responsible for selling “fleet performance” tools – software solutions that use data and analytics to help owners, managers and operators make better decisions on how to manage their fleet. During this time, I have had many different areas of responsibility – front-line sales, product/market expertise and research, matrix leadership and managing teams.

I’m writing this down mainly so I don’t end up forgetting it down the road, but the knowledge and insight is hopefully useful to some.

What is fleet performance?

I have learned to actively dislike the words “fleet performance”. Not because the tools or products are bad, but because It’s a bit of a catch-all phrase that covers everything and nothing. There are tons of different areas of performance you can improve using data analytics and software tools – so what exactly are the vendors of these “fleet performance” solutions really doing?

This brings me to the first thing to look at when getting a market overview – the actual area or thing that they are trying to optimize. I’ve found looking at four main categories makes sense to me:

Fuel consumption – this is the biggest category with the most competition between both existing players and a lot of new market entrants. The investor pitch of “the maritime industry is responsible for 3% of global emissions and we can reduce that by 30% using data, i.e. 1% of global emissions” has been well received in past years, which is why you see the market flooded with new entrants. Whether you can really reduce overall consumption by 30% remains to be seen – personally, I have only seen 5-10% proven in best-case scenarios. It is of course always possible to play with the numbers to show a higher ROI to your stakeholders.

You can largely subdivide the providers in this category into voyage optimization (routes, speeds) and technical optimization (hull & propeller, and machinery maintenance and operation). On top of this, you have some niche vendors looking at e.g., boil-off for LNG carriers, trim, propeller pitch, cargo heating or engine combinations, and the list goes on – these types of solutions target very specific types of ships. It’s important to distinguish exactly what the vendors are doing here from the start – put them in the right box and evaluate from there.

Reliability – here I refer to using the data to improve the uptime of ship equipment to avoid operational disruptions/off-hire or minimize maintenance costs. The big players you see here are generally the equipment manufacturers offering add-on services to an equipment sale, but there are a few independent ones. I am personally skeptical of solutions in this space that aren’t provided by the equipment manufacturer, as AI isn’t perfect yet and you still need intervention/recommendations from trained engineers in many cases. The solutions will get better over time, however, and I may be slightly biased.

Safety – this category is for providers using software and analytics to improve safety onboard. You can also subdivide the solutions in this category; they’re generally either focused on navigational safety by integrating to the bridge (to optimize the passage planning process or using the data to identify unsafe behavior), using augmented reality with cameras (to improve situational awareness), or using CCTV and AI-based object recognition to see if the crew is doing something unsafe elsewhere on the vessel.

Compliance – solutions in this space are generally optimizing a process that shipping companies must do anyway. Some examples would be e-logbooks or sensors integrated to noon reporting for automatic pre-filling, easier ways to maintain data quality for EU MRV/IMO DCS/CII. It could also be having some smart tooling working parallel to existing processes, which alerts or safeguards against possible navigational, environmental or commercial compliance deviations.

Putting this together, you get the beginnings of a vendor comparison matrix:

From this point onwards, the deviation between solutions in the market really begins. I like calling the next column enablers, as these are the minimum requirements for their services to work. Here I find it easiest to divide it into the following levels:

Manual – solutions that are based on manually reported data (e.g., noon reports or logbooks) are the lowest-hanging fruit. When looking at vendors in this space, the key question to ask is how they validate the accuracy of the data.

Integrated – here we find the solutions which are connected to equipment onboard providing sensor data. There are different levels of integration, for example just doing the bridge versus also doing the automation system and engine room. More data isn’t always better – it is important to have a clear purpose for doing the integrations. Once you have sensor data, you still need to solve sensor calibration issues and get an understanding of the impact external factors such as weather have on the vessel. Because of this, integrating vessels isn’t necessarily the solution to poor data quality from manual reporting. Ask the vendors how they are tackling these challenges.

?Cameras – this level uses image analysis to enable some of the prior categories. There is a large variety here for different purposes, either integrating to existing CCTV feeds or installing high-grade marine cameras to enable something else.

Now the second column of the matrix is complete, and you know what the vendor is trying to optimize and have an idea of what data they are using to do it. There is already a huge variety at this point, for example, I have seen camera-based technology being used to solve all 4 categories.?

The next column in the matrix is what I would call service levels, or how the solution is delivered to the end user. There is yet again a massive difference between vendors here, and to me, it makes sense to look at it as follows:

Visualization – these systems simply present the collected data online, usually on a map with some basic graphing tools (that you could do yourself with a BI license and an API). You need to be careful here, because many vendors still use the pitch of “optimizing your fleet performance based on the data”, even though they are not really doing anything with the data. There is a long way to go from having data to making smart decisions based on it. Usually, these solutions are delivered by a shipyard or equipment manufacturer with the main purpose of satisfying certain minimum requirements for a “smart ship” class notation or similar. Many shipowners are now in the situation of having a massive amount of data and not using it for anything – feeling it was a bad investment decision. The reason is that they bought into this “service level”.

Analytics – this is maybe not the best description, but the providers and solutions I would put in this category are the ones analyzing the data with contextual information, and providing the outcome of the analysis through alerts or dashboards/tools. What you need to look at here are two main things; are the dashboards and tools set up to directly solve a business need (or are there extra steps involved, making it cumbersome), and what do you want to be alerted about in real-time versus the logic of the alerts the vendor has in the system. If you’re not careful with the alerts, you could end up with something that looks the same but is actually very different.

Predictive – in addition to the analytics, is the vendor using the historic analysis to provide tools that solve the problems before they arise? How are they doing this, and how does it fit into your existing processes?

Expert-in-the-Loop – this is what I would call the highest service level available in the market. In addition to providing software, a service like this would either be a hotline you can call for advice from experts or outsourcing part of the performance management. A common provider of this would for example be the weather routing providers, or if you have data-driven service agreements with equipment manufacturers. ?

Putting all of this together, we now have a high-level matrix of what to consider when wanting to compare apples to apples in the “fleet performance” space.?

This is the mind map I use when trying to understand the difference between vendors in the market, who generally all use the same marketing language. Some vendors will tick multiple boxes in each column, some only one, but I have found this visual useful when thinking about what they are doing. Hopefully, it helps others as well. I do not envy the owners, operators, and managers who are trying to make sense of all the different options. It’s a really crowded space with all the hype surrounding digitalization.

Mathias