CII Rating Calculation: A Step-by-Step Guide for Fleet Managers

What does CII actually measure — and what does it leave out?

CII (Carbon Intensity Indicator) is an annual, operational measure of how much CO₂ a ship emits per unit of transport work over a calendar year — expressed in grams of CO₂ per cargo-carrying capacity and nautical mile, and converted into an A-to-E letter against IMO reference lines. “Transport work” here is the supply-based proxy G1 defines: the ship’s capacity multiplied by the distance it travelled, not the cargo it actually carried. It does not measure a ship’s design efficiency (that is EEXI’s job) and it does not measure absolute emissions in tonnes: a ship can cut its total CO₂ and still see its letter get worse if the distance it covered fell further, because capacity is a constant and distance is the only term that moves.

The distinction from EEXI matters because the two get quoted interchangeably in commercial conversations. EEXI is a design index — calculated from a ship’s technical particulars, required to come out below its required EEXI, and then not recalculated year on year. It is not frozen for life either: MARPOL Annex VI regulation 23 requires an attained EEXI for each ship and for each ship that has undergone a major conversion. The trigger is that legal threshold, not any change of particulars — a re-engining, a bulbous-bow retrofit or a cargo-system conversion puts the technical file and its verification back on the table if the work meets the Annex VI definition of a major conversion. That determination belongs with class and the flag Administration, and is better settled before the yard period than after it. CII, by contrast, is calculated every year from what the ship actually did, and produces a letter rather than a pass. A ship can pass EEXI comfortably and still attain a D or an E on CII if its trading pattern that year was unfavourable.

Fuel choice does move the CII number, but only on a tank-to-wake basis. The carbon conversion factor (Cf) applied to each fuel type is what carries fuel carbon content into the calculation, so a ship running a low-carbon fuel can attain a better rating than the same ship on residual fuel — while the emissions from producing and delivering that fuel sit outside CII entirely. That boundary is the one to hold on to whenever a CII result is quoted next to another regime’s number: a well-to-wake accounting of the same bunker is answering a different question, and neither figure converts into the other.

The operational corollary matters more than the formula, and it depends on separating two cases that get run together. A ship that spent much of the year at anchor or alongside burns fuel — auxiliaries, boilers, cargo and ballast pumps — while accumulating no distance at all, and that is unmitigated damage to the attained value. A ballast passage is a different animal: it produces miles, and under G1 every one of them counts in full in the total distance travelled, so whether the leg helps or hurts turns only on how much fuel went into those miles. Because capacity is fixed at the ship’s DWT whether the holds are full or empty, a ship steaming economically at ballast draught can attain a better value per mile than the same ship on a deep-draught laden passage. Either way the letter can move for reasons that have nothing to do with the condition of the machinery or the hull.

Which ships are in scope (and which are out)

CII applies to ships of 5,000 gross tonnage and above — broadly the same population that reports fuel consumption to the IMO Data Collection System (IMO DCS, Regulation 27 of MARPOL Annex VI) — and only to ships falling in one of the categories listed in MARPOL Annex VI: bulk carrier, combination carrier, containership, cruise passenger ship, gas carrier, general cargo ship, LNG carrier, refrigerated cargo carrier, ro-ro cargo ship, ro-ro cargo ship (vehicle carrier), ro-ro passenger ship and tanker. Ships of 400 GT and above remain inside Chapter 4 for EEXI and SEEMP purposes, but the attained CII and the letter rating start at 5,000 GT.

Two scope points are regularly missed. First, Chapter 4 does not apply to ships engaged solely on voyages within waters under the sovereignty or jurisdiction of the State whose flag the ship is entitled to fly — so a purely domestic trader sits outside the Chapter 4 rating obligation, even above 5,000 GT. That is not the same as nothing applying. Each Party is expected to adopt appropriate measures so that such ships are constructed and act in a manner consistent with Chapter 4, so far as is reasonable and practicable, which makes the final answer for any given ship a question of flag law rather than of the Annex alone: confirm what your administration has actually enacted, and check separately whether any regional regime reaches the same voyages. Read the exemption as a change of legal basis, not a blanket pass. Second, a ship that does not fit any of the listed categories is out of scope regardless of tonnage — and regulation 19 names further exclusions, at two different levels, which is the distinction to carry into the discussion. Regulation 19.2.2 excludes whole hull types from the chapter: the provisions of Chapter 4 do not apply to ships not propelled by mechanical means, or to platforms (including FPSOs and FSUs) and drilling rigs regardless of their propulsion. Regulation 19.3 then works regulation by regulation rather than chapter-wide: it disapplies regulations 22, 23, 24 and 25 — the EEDI and EEXI requirements — to ships of non-conventional propulsion, with named exceptions for cruise passenger ships and LNG carriers, while leaving regulation 28 and the CII rating standing for them; and it disapplies regulations 22, 23, 24, 25 and 28 to category A ships as defined in the Polar Code. So “excluded” is not one answer per hull: EEXI, SEEMP/DCS and CII each have to be read against the paragraph that governs them. If your fleet includes an unusual hull, confirm its Annex VI category, check it against those two paragraphs separately, and have the position confirmed by your flag Administration or RO before building a CII budget around it.

Within the in-scope population, the applicable metric differs by ship type. AER (Annual Efficiency Ratio) uses deadweight tonnage as capacity and applies to bulk carriers, tankers, container ships, gas carriers, LNG carriers, general cargo ships, refrigerated cargo carriers and combination carriers. cgDIST uses gross tonnage as capacity and applies to ro-ro cargo ships, vehicle carriers, ro-ro passenger ships and cruise passenger ships. Both are “supply-based” indicators in G1’s own terms — the ship’s capacity is taken as a proxy for the actual mass or volume of the shipment on board, which is why the indicator is blind to how much cargo a voyage actually moved: a part cargo earns no credit, and a full hold earns no allowance. The split is set out in IMO resolution MEPC.352(78), the CII Guidelines (G1); check your ship’s category rather than assuming AER by default.

The calculation, step by step: fuel → CO₂ → attained AER / cgDIST

The arithmetic runs in one direction only — from what was burned to what was travelled:

1. Collect annual fuel consumption by fuel type, in metric tonnes, for the full calendar year, drawn from the data collection plan in SEEMP Part II (bunker delivery notes, flow meter or tank sounding data — whichever method the ship’s approved plan specifies).
2. Convert each fuel type to CO₂ emitted, using the fuel-specific carbon conversion factor (Cf) for that fuel — different fuel grades and alternative fuels carry different Cf values, so a fleet burning mixed fuels cannot use one blended factor. The Cf values are the ones specified in the EEDI calculation guidelines (resolution MEPC.308(73), as amended); for a fuel those guidelines do not cover, G1 requires the factor to come from the fuel oil supplier, supported by documentary evidence. Chasing that evidence after year-end is harder than asking for it at the bunkering.
3. Sum CO₂ across all fuel types to get total annual CO₂ emitted, in grams (the formula works in grams, not tonnes, so unit conversion at this step is a common source of error).
4. Establish the denominator — the ship’s capacity (deadweight tonnage for AER ship types, gross tonnage for cgDIST ship types) multiplied by the total distance travelled in nautical miles, as reported under IMO DCS. For AER ships, capacity is the DWT at summer load draught as certified in the approved stability booklet; for cgDIST ships it is GT calculated under the 1969 Tonnage Convention. Note which distance the Annex means: regulation 2.1.12 defines distance travelled as distance travelled over ground, not the through-water log figure that a noon-report template will often hand you first.
5. Divide: attained AER = total annual CO₂ (grams) ÷ (capacity × distance travelled). The result is the ship’s attained value for that calendar year — a single number, before any letter is assigned.
6. Apply correction factors and voyage adjustments, if any, before the value is used for rating. These are not optional refinements invented by the manager: they are set out in the interim guidelines MEPC.355(78) (G5) and cover specific design and operational circumstances. They must be reported to the verifier together with the aggregated DCS data.

The distance term is where most disputes start. Dt is the total distance travelled over the calendar year as reported under DCS — ballast legs and positioning voyages count in full, and there is no provision for leaving out a leg simply because it did the rating no favours. The only distance that comes out is what G5 allows, documented as a voyage adjustment. Note what travels with that figure in the submission: appendix IX to Annex VI asks for fuel oil consumption by fuel oil type in metric tonnes and the methods used to collect it, distance travelled and hours under way, alongside the ship’s GT, NT, DWT, ice class and engine power output — plus any correction factors and voyage adjustments. A distance total that cannot be reconciled against hours underway and the fuel burned is the first number a verifier will come back on, so run that cross-check in-house before the data leaves the office. The practical consequence: a year of short hauls with heavy port and anchorage time will attain a worse value than the same ship on long laden passages, with identical machinery and identical hull condition.

Worked example: one bulk carrier, one year of data

Take a Handysize bulk carrier of 34,000 DWT that travels 42,000 nautical miles over a calendar year. Its logged HFO and MGO consumption, converted with the applicable Cf values, gives a total of 8,400 tonnes of CO₂.

• Total CO₂ in grams: 8,400 t × 1,000,000 = 8,400,000,000 g
• Denominator: 34,000 DWT × 42,000 nm = 1,428,000,000
• Attained AER = 8,400,000,000 ÷ 1,428,000,000 ≈ 5.88 g CO₂ per deadweight tonne-nautical mile

That 5.88 is the ship’s attained value — the number then compared against its required CII to produce a letter. The reference line constants for bulk carriers and the reduction factor for the reporting year come from G2 and G3 respectively (see the next section); the arithmetic above is the part that does not change from year to year: fuel to CO₂, and CO₂ over capacity-times-distance.

Two things are worth noticing about the structure of that fraction. The attained value is inversely proportional to reported distance, so a distance figure overstated by 5% improves the attained value by just under 5% — which is exactly why verifiers reconcile distance rather than take the annual total on trust. And capacity is a fixed constant for the year: in the basic, uncorrected calculation the ship cannot improve its number by loading more cargo — only by burning less fuel or covering more miles. The rating it is finally given can still move on validly documented G5 correction factors and voyage adjustments, per step 6 above; cargo intake is not among them. For a shipowner that is the uncomfortable part of the design — waiting for a better freight market is, in CII terms, an expense.

Run your own vessel’s numbers in the CII rating estimator.

From attained value to a letter: reference lines, reduction factors, rating bands

Once the attained value is known, it is measured against a reference line — a curve specific to the ship type and its capacity — reduced by a reduction factor (Z). Reference line, reduction factor and band boundaries each sit in their own resolution, and that is the single most useful thing to know when checking someone else’s calculation:

Input	Where it comes from
Reference lines by ship type and capacity	MEPC.353(78), CII Reference Lines Guidelines (G2)
Reduction factor Z, relative to the 2019 reference line, for 2023 through 2030	MEPC.338(76), CII Reduction Factor Guidelines (G3), as amended by MEPC.400(83)
Rating band boundaries, by ship type	MEPC.354(78), CII Rating Guidelines (G4)
Correction factors and voyage adjustments	MEPC.355(78), interim guidelines (G5)

Reference line × (1 − Z/100) gives the required annual operational CII for that ship in that year; the midpoint of the C band corresponds to that required value. The ratio of attained to required then places the ship in one of five bands, A through E, using boundary values that are themselves ship-type-specific.

Note that G3 has already been amended once, by resolution MEPC.400(83) adopted at MEPC 83 as part of phase 1 of the review of the short-term measure, which set Z relative to the 2019 reference line for the years 2023 to 2030. A spreadsheet built on the original 2021 factors will produce confident, wrong letters for the later years. Check which version of G3 your calculation tool is using before you trust its forecast — and note that phase 2 of the review runs from spring 2026 to spring 2028 to develop the SEEMP framework and the metrics further, so treat any multi-year CII projection as a planning estimate rather than a settled obligation.

What is stable is the mechanism: the bar moves every year even if the ship’s operation does not change, which means a ship holding a flat attained AER year over year will drift toward a worse letter over time purely because the required value tightened. Planning against this year’s letter rather than the required value three years out is how offices end up surprised.

What each letter does and does not oblige

Keep the two columns below apart when you brief anyone on a letter. The right-hand column is the regulatory consequence under regulation 28. The left-hand column is suggested internal practice — the Annex attaches no duty to the master or the chief engineer on the strength of a rating, and nothing in it makes a letter a judgement on the machinery.

	Suggested internal response (company practice)	Regulatory consequence (regulation 28)
A or B	Establish what actually produced the result before treating it as proof of efficiency: fuel choice, correction factors, voyage adjustments and the year’s trading mix all move the attained value, so a good letter is not on its own evidence of a good speed profile	No corrective action plan required. Administrations, port authorities and other stakeholders as appropriate are encouraged to provide incentives to ships rated A or B (regulation 28.10) — worth asking what a given port offers
C	Review speed and consumption trends: the required value tightens each year, so a mid-C ship has less margin next year than this	None
D (one year)	Review the attained-value drivers (speed, waiting and port time, the laden/ballast split) and the accuracy of the reported fuel and distance data	No corrective action plan required
D (two consecutive years)	Same review, plus start drafting the plan you will need if the third year goes the same way	No corrective action plan required
D (three consecutive years) or E	Treat the plan as a working plan the office can show evidence of, not a filing	Plan of corrective actions required, reflected in SEEMP Part III and verified before the Statement of Compliance can be issued (regulations 28.7–28.9 and 6.8)

What a D or an E means: the corrective action plan and its timeline

A single D year does not, by itself, trigger a corrective action plan under the IMO framework — it is a D rating for three consecutive years, or a single E rating, that requires the ship’s SEEMP Part III to be revised with a documented corrective action plan showing how the required rating of C or above will be achieved.

A simple decision tree for the office:

• First D this year, previous years were C or better → note it, review the attained-value drivers (speed, waiting and port time, the laden/ballast split and the consumption on each), no mandatory SEEMP revision yet.
• Second consecutive D → same review, but start drafting the corrective action plan now rather than waiting for the trigger — a plan built under time pressure is a weaker plan.
• Third consecutive D, or any E → SEEMP Part III must be updated with a corrective action plan, and the plan needs a credible mechanism rather than an intention. It has to consist of an analysis of why the required CII was not achieved plus a revised implementation plan, and it is aimed at achieving the required annual operational CII for the calendar year two years after the one that produced the inferior rating (data year YYYY → target year YYYY+2).

The calendar is fixed, not flag-specific:

1. The reporting year closes on 31 December.
2. The attained annual operational CII and the rating are reported to the flag Administration or an authorised RO within three months of year end — no later than 31 March — electronically, in the standardised format at Appendix 4 of MEPC.395(82), together with the aggregated DCS data including any correction factors and voyage adjustments.
3. The verifier checks the reported data and the attained CII, and the rating is established.
4. The attained CII and the A-to-E rating are recorded on the DCS Statement of Compliance for fuel oil consumption reporting and operational carbon intensity rating, which is issued no later than five months from the beginning of the calendar year (regulation 6.6.4) and must be kept on board for at least five years (regulation 9.12). Each SoC is valid for the calendar year in which it is issued and for the first five months of the following calendar year, which is what bridges the gap while the new one is being produced. Note that five months is the outside limit for issue, not a scheduled date: where reporting and verification run early the new SoC can be in hand well before the end of May. So the question to ask in, say, April is not which year’s document is on board, but whether a valid SoC covers that date and whether the required run of past Statements is still there. Wherever no corrective action plan has been triggered, the SoC is issued on the strength of the verified data alone — and that includes a first D and a second consecutive D, not only C and above. A D on its own does not hold up the certificate; what holds it up is the plan requirement in step 5.
5. Where a corrective action plan is triggered, the plan must be developed, reflected in the SEEMP and verified before the Statement of Compliance can be issued (regulation 6.8) — and the revised SEEMP has to go to the Administration or its RO for verification preferably together with the CII report and in no case later than one month after it (regulation 28.8). That is the real deadline pressure: the SoC is the document a PSC officer and a vetting inspector will ask to see.

The sequence in step 5 is the one offices underestimate, and the one-month rule is why. The trigger is foreseeable from the office’s own data long before anyone else confirms it — two consecutive D years and a third-year attained value heading the same way — so there is nothing to wait for. A plan begun only once the rating is formally established has roughly four weeks to be written, competes for the same naval architect and the same superintendent as the rest of the year’s work, and is holding up a statutory certificate while it does.

Who owns the number: owner, manager, charterer — and what the charter party says

The CII rating attaches to the ship, but the party best positioned to influence it depends on who controls the variables that drive the attained value. Speed and routing decisions on a time-chartered ship often sit with the charterer; the technical manager controls machinery condition, hull and propeller performance, and the accuracy of data collection.

The regulatory obligation itself sits with the Company, which is not automatically the registered owner. MARPOL Annex VI carries over the ISM Code’s definition: the Company is the owner, or whichever other organization — in practice usually the manager or a bareboat charterer — has assumed responsibility for the operation of the ship from the owner and, on assuming it, agreed to take over the duties the ISM Code imposes. The SEEMP is a Company document in that sense, and IMO’s verification guidance (MEPC.347(78)) is written around administration verification and company audits of SEEMP Part III. In an owner-managed fleet the Company is the owner. Under third-party technical management or a bareboat charter it frequently is not, and the manager who assumed “the owner carries it” may have misread its own exposure. Establish which entity is the declared Company for each hull — the ISM Document of Compliance is the first place to look — before allocating anything contractually. Where the structure is unusual, confirm the position with the flag Administration rather than inferring it: how the burden finally falls as between the parties is settled by the management agreement and the charter, and by flag law, not by the Annex on its own.

That split is the whole problem with CII in a time charter: the regulatory obligation and the operational levers sit with different parties. The Company carries the SEEMP and the corrective action plan; the charterer orders the speed, the waiting and the ballast legs that drive the attained value. Standard clause wording for time charter parties circulates to address that gap, but a clause is only as good as its edition and its fit to the fixture: establish which body published the wording you are being offered and which version it is, and have it reviewed as a legal document for the specific charter rather than adopted as a template — none of this is legal advice, and no two charter parties allocate the same risk. Whatever the source, the questions any such clause has to answer are the same three: who may instruct what speed, what consumption and distance data each side must share and how often, and what happens to hire or to an indemnity if charterer-directed employment pushes the ship below its agreed rating.

Keep the two kinds of consequence distinct when you brief management. What a poor rating obliges the Company to do — revise SEEMP Part III, get it verified before the Statement of Compliance issues — is a regulatory requirement. What a charterer, a financier or a vetting department does with a D is commercial pressure.

What the letter itself is not is a tariff. What regulation 28 attaches to a D or an E is a plan, not a payment: the regulation requires calculation, reporting, verification and — when the three-D or single-E trigger is met — a plan of corrective actions that the ship then duly undertakes, and it stops there. The review mandated by regulation 28.11 itself put “the need for enhancement of the enforcement mechanism” on the agenda, which is a fair description of where the regime stands. Do not let that harden into “CII cannot be enforced”, and do not read the absence of an IMO tariff as the absence of consequences. The consequence the Annex does attach is the certificate: a ship rated D for three consecutive years or rated as E shall not be issued a Statement of Compliance unless the plan of corrective actions is developed, reflected in the SEEMP and verified (regulation 6.8). A port State inspection may verify that there is a valid Statement of Compliance, an International Energy Efficiency Certificate and a SEEMP on board, and may inspect whether the SEEMP is being duly implemented (regulations 10.5 and 10.6). Where there is sufficient evidence of a violation of the Annex, it is the flag Administration that causes proceedings to be taken, and it does so in accordance with its own law (regulation 11.4) — which is where the actual exposure has to be checked: what your flag has enacted, and what the administrations in your trading areas have. So: no tariff on the letter, real consequences attached to the certificate, national law behind both, and commercial pressure from everyone who reads the letter.

The data discipline behind the rating: what verifiers ask for

Because the entire calculation runs from fuel and distance figures the ship itself reports, verifiers focus less on the arithmetic (which is mechanical) and more on whether the underlying data is defensible:

• Consistency of the data collection method with what the approved SEEMP Part II data collection plan specifies. The method is recorded in a plan the Administration or its RO has confirmed as compliant — for SEEMP Part II a confirmation of compliance is issued before data collection starts and is retained on board (regulation 5.4.5) — so a ship that has to move from flow-meter to bunker-delivery-note figures mid-year should have the change carried into the plan and the reason recorded, rather than letting the switch surface for the first time in the annual aggregate. That original confirmation is not an approval of every later change: confirm the notification and amendment route with your flag or authorised RO before the method changes, not at verification.
• Traceability of fuel figures back to bunker delivery notes, tank soundings, or flow meter logs, not office estimates. A ship whose bunker survey records are tidy has already done most of this work; one that settles quantity disputes informally has not.
• Distance figures reconciled against a voyage-by-voyage trail — noon reports, deck and engine log abstracts, and the company’s retained electronic reporting dataset — rather than a single aggregated annual number with nothing behind it. This is not housekeeping, it is a requirement: the disaggregated data underlying the appendix IX figures has to be readily accessible for not less than 12 months from the end of the calendar year it covers, and made available to the Administration on request (regulation 27.8). Do not assume VDR output will fill that gap: the VDR is a safety recorder working to its own, far shorter retention regime, and what an individual installation holds and can actually export varies by system and by company procedure. The twelve-month obligation is the company’s to meet, with separate, accessible voyage and distance records standing behind the annual figures.
• Treatment of ballast legs, waiting time and port time. Waiting at anchor and port time add fuel to the numerator and nothing to the denominator, and that is the pattern that does the real damage to the attained value. Ballast legs are not the same problem: they bring their own miles with them, and per mile they are neutral to favourable where the ship is steaming economically. Verifiers expect the manager to be able to explain the year’s employment pattern and how it moved the number, not just report the aggregate.
• Correction factors and voyage adjustments claimed under G5, each with the evidence that supports it. A voyage adjustment asserted without a documented basis is the quickest way to turn a routine verification into a long one.
• A mid-year change of flag or of company. On a transfer between Administrations the ship reports the aggregated data for the portion of the calendar year belonging to the losing Administration, and on a change from one company to another the aggregated data for that company’s portion — in both cases on the day the change completes or as close to it as practical, with the disaggregated data on request (regulations 27.4 and 27.5). The indicator itself is not pro-rated: in a transfer year the ship still calculates and reports an attained annual operational CII for the full 1 January to 31 December period (regulation 28.3). A handover pack that stops at bunker ROB and the certificate folder leaves the incoming manager to rebuild someone else’s fuel and distance data months later; ask for the DCS dataset and the data collection methodology at takeover, not in March.
• SEEMP Part III itself, which has to carry the ship’s performance rating for the previous three years and is subject to verification and company audits — so it is not a document that can be filed once and left. It also has to state the required annual operational CII for the next three years and the implementation plan documenting how that will be achieved (regulation 26.3.1), which is where a forecast built on the wrong version of G3 shows up in a document a surveyor is reading. The three-year history on board is also the first place a vetting inspector can see a D trend that the current year’s letter alone would not reveal.

Three mistakes worth naming directly, because they recur:

• The frozen spreadsheet. A team builds the attained-value model once and reuses it, carrying forward last year’s reduction factor, Cf values or capacity figure. For a ship that has been converted — or after an amendment like MEPC.400(83) — this silently produces a wrong number that surfaces at verification.
• Capacity taken from the wrong document. DWT for AER must be the summer load draught figure from the approved stability booklet, not a commercial deadweight quoted in a fixture recap or a vessel particulars sheet. A few hundred tonnes of difference moves the attained value.
• Rating management left to the fourth quarter. The attained value aggregates a full twelve months, so a measure taken in November touches only the last weeks of the averaging period, while the same measure in January would have shaped the whole year. The measures IMO itself names are hull cleaning to reduce drag, speed and routeing optimisation, low-energy lighting and solar or wind auxiliary power for accommodation services; propeller polishing belongs on any practical list alongside them. Every one of them rewards being applied early. Monitoring a running estimate of the attained value month by month, rather than waiting for the verified letter, is what turns CII from a reporting exercise into something a superintendent can actually manage — and if the ship’s DCS data already goes to class, there is a fair chance a preliminary attained value and rating is already being computed from those uploads — DNV, for one, calculates it for its DCS customers and exposes a running fleet rating on its own dashboard, and other ROs offer comparable tools. Check what your provider’s service actually covers; where it is there, ask for it quarterly instead of meeting the letter for the first time in March.

Related reading

• EU ETS shipping compliance guide — a separate carbon-compliance obligation, triggered by voyages to, from and between ports under the jurisdiction of an EU Member State rather than by flag or by MARPOL applicability; work out which of your ships it actually reaches instead of assuming it tracks the CII fleet.
• FuelEU Maritime compliance strategy — a separate fuel-intensity regime; do not conflate its mechanics with CII’s.
• IMO Net-Zero Framework guide — the separate, longer-term IMO workstream; CII is the short-term measure and is not part of it.
• Regulations affecting shipowners in 2026 — calendar and context for this and other obligations.
• Existing fleet maritime tech frontier — efficiency improvement options for ships already trading.
• Spreadsheets vs. evidence-based fleet oversight — the data discipline this guide’s last section depends on, in more depth.