Modern shipping increasingly relies on data to understand vessel performance and fuel efficiency. Sensors, onboard monitoring systems and analytical models now generate vast amounts of operational information. Yet the reliability of any efficiency claim ultimately depends on the quality of that data. This insight focuses on why measurement accuracy, sensor calibration and data integrity are critical for verifying small efficiency gains.
Data quality is the foundation of credible performance claims. Without rigorous calibration, validation, and governance, even the most advanced AI or physics-based models will amplify uncertainty instead of reducing it.
If instruments have an error margin of ±10%, it becomes near impossible to prove a 5% or smaller gain. Small improvements disappear in the noise. For instance, a torque meter accurate to ±10% cannot reliably detect a 5% efficiency increase.
As Vincent Joly, Smartship Manager at Bureau Veritas explained, “You cannot measure below one millimetre when you are trying to measure a difference of microns.”
Nicolas Bathfield, Project Manager at Stena Teknik, and Sébastien Roche, General Manager, Shipping Performance & Innovation at TotalEnergies both warned that when efficiency gains are small, they can easily disappear in the noise of environmental variation.
Even with good data, weather-driven variability can obscure marginal effects.
Sensor and instrumentation limitations add further challenges. Many operators still rely on noon reports, which are manually logged once per day and prone to approximation and crew bias. These low-frequency datasets cannot detect marginal efficiency improvements.
High-frequency data logged every few seconds improves visibility but it is not a silver bullet. Sensor drift, electrical noise, and calibration errors can distort measurements.
Calibration errors in torque meters alone can shift calculated efficiency by several percentage points, more than enough to invalidate 3–5% fuel-saving claims.
Connectivity gaps add another issue. Data streaming is often interrupted during heavy seas, meaning key performance periods are lost.
Artificial intelligence models and digital optimisation tools are also dependent on the quality of the data they receive. If the underlying data are noisy or incomplete, even sophisticated optimisation systems will produce unreliable or misleading results.
For efficiency gains in the 5–10% range to be credible, models must resolve uncertainty to within roughly 1–5%. Otherwise the benefits disappear into the noise of measurement errors and environmental variability.
Download ‘The 5-10% Illusion’ and explore why the maritime industry’s reliance on unverified efficiency claims is eroding trust, and discover a framework to restore measurement discipline.
