by Niclas Priess*
“If a man knows not to which port he sails, no wind is favorable.” — Seneca
Although this quote is two thousand years old, its relevance in the context of contemporary business decisions and related data consumption is undeniable. Growth in data capture from increases in the number of connected devices and sensors has led to an explosion in the size of data sets and a resulting myriad of data, analytics, and machine learning/AI platforms. Today, there is nothing novel in stating that data sets are growing and that there is no shortage of evidence that data has become paramount to business success.
Yet, in the face of the exponential growth of data and broader selection of tools for analytics, it can be a real challenge to focus attention and investments when faced with several data and analytics providers that, on the surface, may appear identical in their offering. This post will try to establish a set of guidelines for shipping companies and their executive teams in driving successful data initiatives in their organizations and subsequently explores the role and responsibility of data providers to build offerings that live up to the promises of modern technology.
For the tramp shipping industry, the increased data capture from onboard sensors and Automatic Identification Systems (AIS) has already led to more transparent and efficient operations, yet we are only seeing the beginning of what is possible for the shipping industry with the right data and analytics tools. Let us dive into the perspectives of how shipping companies position themselves best to deliver on the promise of data-driven decisions.
How do shipping companies plant the seeds for successful data-driven decision-making?
- Dedicated team for driving the data agenda:
Returning to the Seneca quote above, if a company does not have a clear direction for what it wants to achieve with data, no dataset and analytics tools will suffice. No shipping executive will deny that data and digitization is at the forefront of the corporate agenda, yet being on the corporate agenda is only the first step in ensuring success with data.
The shipping companies that are pioneering data and analytics for decision-making and business operations have dedicated teams that ensure implementation and support the data initiatives set out by top management. The key responsibilities of these teams are i) concretizing the data agenda into corresponding data needs, ii) defining and amending technology team resources, and iii) evaluating and selecting data suppliers, but most importantly these teams ensure internal user adoption. The internal focus on user adoption is particularly important to break old habits and to ensure that investments in new technologies, such as software and data, do not go overlooked by actual users in the organization.
- Embrace existing automation opportunities:
Software and data is not replacing all the people in the shipping industry just yet, but it is already making chartering managers, operation managers, research analysts, and data science teams more efficient in their day-to-day routines.
Shipping companies should leverage existing automation opportunities brought by modern technology such as APIs[1] and SDKs[2]. As a concrete example, any data that lives in a spreadsheet is probably being manually updated and this process could be automatized by leveraging an appropriate technology.
Embracing automation faces the challenge of upskilling the existing teams with new tools, e.g. Python[3]. Modern datasets have outgrown the capabilities of spreadsheet tools. With a corporate agenda of digitization, the pioneers will choose to invest in employees to learn fundamental building blocks of modern data science (e.g. learning Python) to work with data at scale. Heuristically, managers can think about upskilling employees as moving up the “Data Adoption Ladder” as depicted in Figure 1.
Figure 1 - Data Adoption Ladder
- Avoid swiss army knife solutions:
With the rise of cloud computing over the past decades, microservices and modularized software packages have become the norm, each focusing on functionality that’s done particularly well. Modern software and data companies are built in a similar way, often focusing on delivering a narrowly defined service.
This implies that the modern shipping technology stack will consist not just of a few, but of several data and software providers that each excel in a narrowly defined topic. Yesteryears’ shipping software providers promise to do all of these things well, but often only deliver average performance in any one area.
- Data discoveries can be scary:
Shipping is an industry where long-standing relationships drive business decisions. Data is not here to change that dynamic, but adopting more data in decision-making can lead to surprises when shedding light on information that has lived under the mutual trust of these relationships.
Whether data sheds light on previously opaque trades or vessel underperformance, shipping professionals should embrace the fact that transparency is here to stay and act on the basis of the latest set of information.
- Data by itself is not a competitive advantage:
This is somewhat a controversial statement, at least at the time of writing. Historically, tramp shipping and the related commodity markets have been notoriously opaque, but, to quote Stewart Brand, “information wants to be free" and it would be foolish to expect that the historical opaqueness of these markets will continue to persist.
Shipping companies should recognize that data in itself is not a competitive advantage if the organization is not able to interpret, analyse and act on the data. Data analytics at scale are being democratized by cheap cloud computing, open source software (e.g. Python, R), and machine learning (e.g. TensorFlow[4]), allowing every willing organization to boost its data analysis capabilities.
Shipping companies are not solely responsible for driving data-driven initiatives, much of the burden lies on the providers of data. The following summarizes the responsibilities of the data providers for the tramp shipping industry in the context of the driving data-driven decision-making going forward.
The responsibilities of modern data providers for the tramp shipping industry.
- Leverage modern interfaces and delivery:
Historically, data providers have been trying to capitalize on their datasets by maintaining them behind strict conditions for usage often enclosing them in a proprietary platform making it near impossible for users to access the data outside of the platform perimeters.
In the current era of specialized data providers, cloud computing and open source analytics software, enclosing data in a proprietary platform will only cause user frustration. Today, no single dataset encapsulates all the truth and being able to do further data analysis, manipulation and visualization is where the competitive advantage lies.
Modern data providers should not let their data live in a closed and proprietary ecosystem, but have a responsibility to deliver data through the interfaces that the users already spend their time in. Consuming data should be friction-less for the user irrespective of interface of delivery.
- Openness and transparency:
Datasets that live in static forms (e.g. spreadsheets or PDF documents) or enclosed data platforms (as described above) often lack proper documentation on the methodology for collecting, combining or calculating the dataset. This is problematic today as the data in itself is of limited value, but will typically serve as one of several inputs for an analytical model.
Modern data providers should have a transparent methodology that provides users with confidence around the information they are consuming. Users should be able to leverage the data in combination with other data sources.
- Join keys are king:
Auren Hoffmann is a thought leader within data-as-a-service businesses and a central idea of his work is that data in itself is not very valuable, but that its value increases as a function of the possibilities to join it with other data. Auren summarizes this nicely on Twitter:
In order to join two datasets a common key is needed, this key is referred to as a join key. In shipping, examples of join keys are: a vessel’s imo number or a port’s UN/LOCODE.
Join keys are king in data science and modern data providers should provide datasets with join keys that allow their users to combine multiple datasets to get a holistic picture of the problem they are analyzing.
- Data rots quickly when markets move fast:
Tramp shipping markets are complex and fast-paced, meaning that temporal opportunities may arise as a consequence of imbalances in supply and demand. Capitalizing on these opportunities requires data in a timely manner. Today, much information still lives in emails, spreadsheets, and PDF documents. It is no surprise that these formats often do not capture the temporal opportunities in a timely manner.
Data rots quickly in tramp shipping and, therefore, modern data providers should minimize the time between when a new piece of information is available to the provider and when it is available to the user.
We are only at the beginning of democratizing data and analytics in the tramp shipping markets and emerging technologies such as VHF Data Exchange System (i.e. the next generation of AIS data), autonomous vessels, and blockchain for commercial contracts will increasingly contribute to new datasets that will aid tramp shipping companies in driving their future decisions.
That is if they adopt appropriate frameworks for data analysis and engage with the new and next generation of data and software providers. The data providers in turn, have their role to play for delivering data to modern standards.
[1] Application Programming Interface: an interface that defines interactions between multiple software applications or mixed hardware-software intermediaries.
[2] Software Development Kit: a collection of software development tools in one installable package.
[3] Python is an interpreted, high-level and general-purpose programming language.
[4] TensorFlow is an end-to-end open source platform for machine learning.
*Niclas Priess is the co-founder of Oceanbolt. Oceanbolt is a data-as-a-service company leveraging geospatial analytics to provide real-time intelligence for the dry bulk markets. Learn more about Oceanbolt on its website: https://www.oceanbolt.com/