Data Is a Fleet Asset

Ask a superintendent what a main engine is worth and you will get a number, a maintenance plan, and the name of the person who answers for its condition. Ask what the fleet's data is worth and you will usually get a pause, then, after some thought, an IT budget line. The pause is the subject of this course.

An asset, in the accountant's sense

Calling something an asset makes three claims: it cost money to acquire, it produces returns, and it degrades unless maintained. Fleet data passes all three tests without straining.

It costs plenty to acquire: sensors and their integration, connectivity, and the least visible item on the invoice: thousands of officer-hours a year spent compiling reports. It produces returns in exactly one way: when a decision is made because of it. A terabyte that never changes anyone's mind has the same return as an empty disk. And it degrades: sensors drift out of calibration, an unrecorded equipment change quietly breaks a time series, the analyst who knew why one year looks odd moves on, and what was knowledge becomes, once again, just numbers.

The framing has teeth because of what it forces you to assign. Every other asset in the company has an owner, a quality standard and a maintenance plan. Data usually has none of the three, which is precisely why it underperforms.

The measured immaturity

That this is the industry's general condition is not an opinion; it has been measured. Lloyd's Register and OneOcean's digital-transformation research: the report Mastering maritime data for competitive advantage is open, benchmarked the sector's digital maturity at roughly 2.1 to 2.2 on a four-point scale (2.1 on the launch page, 2.23 in the report itself), and its data standardisation at about 2.45 of 4. Halfway up the ladder, in other words: the typical fleet collects far more than it can use, and no two of its ships describe the same event in quite the same way.

Read those figures the way you would read a competitor analysis, because that is what they are. If average data competence across the industry sits mid-scale, then data competence is not table stakes: it is differentiation, available at the price of discipline. The physics of your ships is public knowledge; what your fleet can know about its own operation is not.

What unusable data costs

Unusable is not the same as absent. Unusable data gets collected, transmitted, stored and dutifully backed up, and then fails at the moment of use. The costs are easy to recognise once named:

Analyst time. Ask how the working week divides between cleaning data and analysing it. In most fleets the honest answer is embarrassing: skilled people employed as manual filters.
Every question becomes a project. When formats differ by ship and meaning differs by officer, each new question means re-assembling the dataset from scratch. Curiosity acquires a lead time measured in weeks, so fewer questions get asked.
Decisions default to anecdote. When the data cannot answer within the meeting, seniority answers instead, and the fleet's most expensive habits survive another budget round.
Disputes are lost. Charter-party performance claims stand or fall on the quality of the records behind them; so do warranty claims against coating and equipment makers.

Quality decides what you can know

The deepest cost is subtler. A measurement system does not merely describe the operation: it bounds what the company can ever know about it. If the daily fuel figure carries five per cent of uncertainty, a two per cent improvement is not merely hard to see; it is invisible, however sincerely everyone stares at the trend line. And the knowledge cannot be bought back later at any price, because the period it described is over. Data collected without units, timestamps or context is not an archive; it is noise with a filename.

Measurement bounds knowledge: the claimed saving sits entirely inside the uncertainty of the number it would have to be read from. The values are indicative, but the ordering is the point, no amount of staring at a trend line recovers a signal smaller than its own noise.

The five and two per cent above are round numbers, chosen to make the geometry legible. Real bands are rarely that tidy, and they are not only the instrument's. On one of the two sister VLCCs of about 317,000 DWT whose one-minute records lesson 3 draws on, the monthly mean deviation of shaft power from the corrected model-test baseline the two sisters share reads 2.4 per cent in July 2024 and 40.2 per cent in May 2023, on 5,000 and 12,955 analysis-ready minutes respectively. Weather, loading and hull condition all move that indicator, and not one of them holds still while an efficiency project runs. Two per cent of the fuel bill is real money on a ship that size; separating it from a band of that width is not a matter of staring harder at the chart. It is a matter of measuring better and comparing like with like, which is most of what the rest of this course is about.

That is why this course talks about quality long before it talks about platforms. Dashboards are the visible tip. What a fleet can know is decided far below the waterline, at the point of collection.

Where this course goes

Lesson 2 maps the onboard landscape: what noon reports and high-frequency autologging can and cannot each support. Lesson 3 is the craft: validation, outliers, units, timestamps, gaps, versioned pipelines. Lesson 4 covers what the regulators did to fuel data: a burden that is also, handled properly, a gift. Lesson 5 looks outward through AIS; lesson 6 borrows the class societies' checklists for what a monitoring installation must actually do; lesson 7 assembles the architecture and prices build against buy; lesson 8 connects the whole pipeline to decisions: the only place data earns anything.

One habit to adopt from today: when a number is shown in a meeting, ask whether a second analyst, given only the stored data, could reproduce it. If the answer is no, it is not fleet knowledge yet. It is an anecdote with formatting.