TL;DR: A voyage saving is only worth what it survives. Most are claimed against a plan that no longer exists, or against a fleet average that does not describe the vessel, so finance discounts them on sight. A saving holds up when the counterfactual comes from the vessel's own fuel model and the voyage it describes reconciles to the same events the commercial system booked. ZeroNorth's vessel-specific fuel models achieve 93.8% median accuracy against measured consumption.
Operations reports a saving on the voyage. The commercial system shows a different voyage. Both are describing the same ship on the same passage, and nobody can say which version is the one the business should book.
This is rarely error and never fraud. The saving and the voyage record were simply assembled from different source events, and a number that cannot be reconciled does not reach a profit and loss account. It gets filed as an operations claim instead.
Which turns the buying question around. When you evaluate voyage optimisation, the thing to test is not how much fuel the vendor says it saves. It is whether the saving it reports can be set beside the voyage your voyage management system already holds, and still stand.
Why a voyage saving loses credibility on the way to finance
A voyage saving loses credibility when nobody downstream can reproduce it from the voyage record the business actually books on.
A saving is a difference between what happened and what would otherwise have happened. For anyone outside operations to accept it, two things have to hold at once. The counterfactual has to be believable for this specific vessel. And the voyage being described has to be the voyage the commercial team recognises: same departure, same arrival, same consumption, same dates.
The second condition is the one that quietly fails. If the arrival report that updates the commercial record is not the same arrival report the optimisation was measured against, the two versions of the voyage will never agree, however good the routing was. The saving and the booked voyage are then describing different events, and no amount of further optimisation will close the gap.
What makes a voyage saving defensible
A saving is defensible when the counterfactual is modelled from the vessel's own performance and the consumption behind it was validated before it was counted, not after it was questioned.
That is a mechanical claim, not a marketing one, and it is worth taking apart. ZeroNorth voyage optimisation is built so that each of these holds:
The counterfactual comes from a vessel-specific fuel model rather than a class average. This matters commercially because a baseline borrowed from a similar hull describes a ship you are not operating, and a charterer only has to point that out once for the whole figure to be set aside.
That model is scored against what vessels actually burned, not against itself. ZeroNorth's vessel-specific fuel models achieve 93.8% median accuracy against measured consumption. A model that has never been scored against measured outcomes is an assumption rather than a baseline, so the figure and the method behind it are fair things to ask of any vendor on your list, including us.
Consumption is validated on the way in. Figures that fail a sanity check are flagged before they enter the savings calculation rather than after someone downstream notices the total has moved, which is the difference between a correction and a credibility problem.
The models sit on more than 2.7 million optimised voyages, so the expected behaviour a vessel is measured against is drawn from observed operations at scale rather than from a handful of comparable passages.
Each of those is doing the same job: making the number survive its first contact with someone who has a reason to doubt it.
Three ways to claim a voyage saving, and what each one survives
The three common approaches differ in what they compare against, and that choice decides who will accept the result.
What it looks like when the record reconciles
When the voyage record reconciles, the saving stops being an operations report and becomes a commercial input.
TMA Bulk is the clearest example of the shift on ZeroNorth's own deployments. Read how TMA Bulk turns voyage planning into commercial optimisation: the point of the story is not that the routing improved, it is that the optimisation output became something the commercial desk could act on, rather than a report the technical team circulated and the commercial team worked around.
That is the test worth applying to your own fleet. If the figure only ever travels in one direction, out of operations and into a slide, it is not yet a commercial input. If someone on the chartering desk can open the voyage, see the same consumption operations saw, and price the next fixture against it, then it is.
The bottom line
Ask any vendor the same three vendor-neutral questions before you ask about price.
What is the saving measured against, and is that counterfactual specific to this vessel? What accuracy does that model achieve against measured consumption, and by what method? And at what point in the flow is consumption validated, before the figure is calculated or after someone queries it?
For ZeroNorth the answers are specific. The counterfactual is a per-vessel fuel model, not a class average. Those models achieve 93.8% median accuracy against measured consumption. Validation happens on ingest, before the calculation rather than after a query. And the models are built on more than 2.7 million optimised voyages. If you want to go further into how platforms differ on this, we have written separately on choosing a voyage optimisation platform and on what good voyage planning looks like in practice.
See how voyage optimisation handles it, review pricing, or book a demo and bring one of your own completed voyages to the call. Reconciling a real voyage on screen settles the question faster than any deck will.
