The Company on Paper and the Company Your Crew Actually Works For

Every HSEQ manager knows the feeling of an audit that goes well and a fleet that does not. The SMS is current, the drill matrix is complete, the near-miss numbers look healthy on the monthly report, the last two PSC calls closed with a couple of observations. Then something happens on a ship where nothing was supposed to happen, and the investigation surfaces a version of the company nobody wrote down: the Chief Officer who knew the procedure did not work in that configuration and had stopped saying so, the AB who saw the step being skipped and assumed everyone did it that way, the report that was never filed because the last person who filed one spent a month explaining himself.

None of that is visible in an audit, because an audit inspects the system as documented and as demonstrated on the day. It answers whether the company has a procedure, whether the crew can produce it, whether the records exist. It cannot answer the question that actually predicts behaviour, which is what the crew believes will happen to them if they follow that procedure when it costs the vessel four hours.

That belief is what organisational climate means. It is the shared, unwritten answer to “what is it really like to work here, and what does this company really reward.” Perception drives behaviour far more reliably than policy does. A crew convinced that reporting a near miss invites blame will stop reporting near misses, and your incident statistics will improve for exactly the wrong reason. We built Soundings to measure that layer, because it is measurable, and because measuring it well is harder than most people expect.

Why asking directly does not work

The obvious approach is to ask. Send out a staff survey, put “How is the safety culture on board?” on it, add a comment box, count the answers.

This produces material you cannot use. Broad questions invite broad answers, and broad answers cannot be compared to anything. If 62 percent of respondents agree that “safety culture is good,” you have learned nothing about which part of the organisation is failing, whether 62 percent is high or low for a shipping company, or whether it is better or worse than last year, since last year’s question was phrased slightly differently. Worse, home-made questions carry hidden assumptions. Ask “Does management support you?” and every respondent supplies a private definition of support: the Master is thinking of the DPA answering his phone at 0300, the Bosun is thinking of manning levels, the Second Engineer is thinking of whether anyone said thank you.

Then there is the boardroom problem. A number invented in-house is a number that can be argued away in-house. Bring a home-made 62 percent to a management meeting where somebody would rather not fund another deck rating, and it becomes a conversation about the survey rather than about the fleet.

Serious measurement of organisational climate has existed in research for decades, and it works on a different principle.

The instrument: 82 concrete statements

The core of the Soundings survey is the Organizational Climate Measure, developed by Patterson and colleagues and published in 2005 after validation across a large sample of organisations. It has been used and tested by researchers ever since, which matters for reasons we will come back to.

The OCM asks nothing abstract. It presents 82 short statements about ordinary working life and asks the respondent how true each one is, on a four-point scale from “Definitely false” to “Definitely true.” There is deliberately no neutral middle: every answer commits, gently, one way or the other. Statements like “People here are always looking for new ways of looking at problems.” Or “Management involve people when decisions are made that affect them.” Or “People here feel under pressure to meet targets.” Nobody has to define a concept. They only have to recognise their own working week.

The engineering sits in how the statements are grouped. Each underlying dimension is measured by a bundle of four to six statements that approach the same thing from different angles. It works the way you take soundings: one reading can be off for a dozen innocent reasons, but three consistent readings in the same tank tell you something real. If a crew rates every statement in the supervisory support bundle low, that pattern is not noise and it is not one person having a bad trip.

There are 17 such dimensions. Autonomy, supervisory support, training, pressure to produce, clarity of organisational goals, innovation and flexibility, performance feedback, welfare, participation, effort, efficiency, formalisation, tradition, reflexivity, integration, outward focus, quality. Negatively worded statements are reversed when scored, so that every dimension ends up pointing the same way on a scale of 1 to 4, and higher is always better. You never have to remember which direction a number runs.

A dimension score is not an audit finding. It is not a fact about the company in the sense that a valid certificate is a fact. It is the crew’s shared perception, and that is exactly its value, because perception is what governs whether a procedure gets followed at 0400 in bad weather with a charterer waiting.

Because the instrument is published and validated, the numbers behave. Everyone answers the same statements, so a score means the same thing on your VLCC as on your PSV, in your Manila pool as in your Gdynia pool, this year as next year. And when the results go to a board, they arrive attached to a research instrument rather than to somebody’s opinion about what to ask.

The map: four competing values

Seventeen numbers is more than any management team will hold in their heads. The Competing Values Framework is the map that arranges them into something a director can read in one glance.

It works on two tensions that every organisation lives with. The first is where attention points: inward at your own people and processes, or outward at the market, the charterers, the competition, the regulatory horizon. The second is how the organisation holds itself together: through flexibility and adaptation, or through control and stability. Cross those two, and you get four quadrants.

The inward and flexible quadrant is about People: welfare, participation, training, supervisors who back their crew. The inward and controlled quadrant is about Rules: procedures, stability, doing things the way the company has always done them. This is the world of the SMS manual, the permit to work, the checklist. The outward and flexible quadrant is about Change: new ideas, noticing that the industry has moved, adjusting how you work before you are forced to. The outward and controlled quadrant is about Results: efficiency, targets, deadlines, performance feedback, delivering what was promised.

The word “competing” is the point of the whole framework. These four pulls genuinely trade against each other. Effort spent tightening procedures is not spent loosening them for local judgement. Attention spent hitting the schedule is not spent on welfare. No organisation maximises all four, and every organisation therefore gives something up. The only question is whether the trade was chosen deliberately or arrived at by accident.

Shipping companies tend to produce a recognisable shape: strong on Rules, strong on Results, weaker on People, weakest on Change. Read as a sentence, that profile says “a procedural and demanding company that cares somewhat less about its crews and is not particularly interested in their ideas.” Sometimes that is precisely the intended shape. A tanker operator with a demanding oil major vetting programme should score high on Rules; low Rules scores in that trade would be alarming. But the weak Change corner has a cost that shows up on the safety side, because a crew that has concluded nobody upstairs wants to hear their ideas is a crew that has also concluded there is no point mentioning the near miss, the awkward valve arrangement, or the procedure that only works alongside in calm weather.

The four-quadrant profile is, in our experience, the one picture that lands in a boardroom without a statistics lesson.

The maritime layer: safety climate

The OCM was built for organisations in general. Seafaring has its own failure modes, so we added a section of 12 statements written for the people who work on ships, scored the same way and reported alongside the 17 dimensions.

Safety climate is the crew’s shared answer to a single question: what does this company actually value when safety collides with the schedule? The statements deliberately target the seams where paper and practice come apart.

Management commitment is the baseline, tested by whether the company emphasises safety in everyday work rather than in the annual message from the CEO. Compliance when nobody is watching is tested from both directions, by asking whether crew follow safety procedures when no one is checking, and separately whether the respondent sees colleagues skipping safety steps to get the job done. People are often more honest about what they observe than about what they do.

System responsiveness is where a living SMS separates itself from shelf-ware. Two statements carry that weight: whether something is actually done when a crew member reports a safety problem, and whether a procedure gets changed when the crew says it does not work in practice. A company that scores low on the second one has a manual that has stopped learning, and every seafarer on board knows it long before the auditors do.

Commercial pressure gets asked about plainly, because the crew experiences it plainly: whether safety comes second when the schedule is tight, and whether the workload makes it hard to follow safety procedures at all. Preparedness asks whether people genuinely know what to do in a fire, a man overboard, or an order to abandon ship, which is a different question from whether the drills were logged.

And then there is the one we would keep if we could keep only one: whether you can report a mistake or a near miss without fear of being blamed. Just culture in a single line. If that item scores low, the rest of your safety reporting data is fiction, and no amount of analysis further downstream will fix it.

Turning scores into a short list

Results arrive as a ranked table, and the useful end is the bottom. The two or three lowest dimensions are your action list. Everything else is context.

Raw scores only become meaningful through comparison, and there are three that matter. The first is internal: how a dimension ranks against your own other dimensions, which tells you where this company is weakest relative to itself. The second is the industry benchmark, which is what keeps you from chasing the wrong problem. Autonomy at 3.1 looks mediocre until you see the maritime pool sitting at 2.8, and remember that seafaring is a low-autonomy occupation by design; a Third Officer does not get to reinvent the passage plan. The third comparison is your own history. Trend over successive surveys is the honest test of whether last year’s intervention did anything, and it is the number that is hardest to argue with, because the instrument did not change.

Fleet averages hide more than they show, so the breakdowns are where the real findings live. Supervisory support at 3.1 across the fleet looks unremarkable. Split it by rank and find officers at 3.6 and ratings at 2.4, and you have located an actual problem with an actual address. The same applies to nationality, department and vessel type. Anonymity is protected by suppression: no breakdown is displayed for a group smaller than five respondents, because with three people in a category somebody will eventually work out who said what, and once a crew believes that is possible, the survey is finished as a source of truth.

The strongest material comes from cross-referencing the two halves of the instrument. Suppose Pressure to Produce scores badly among the 17 dimensions, and separately the safety climate statement about the schedule pushing safety into second place also scores badly. Individually each is a data point. Together they are evidence that commercial pressure is eroding safety compliance in this fleet, expressed in the crew’s collective voice, in a validated instrument, with a benchmark beside it. That is an argument a board has to answer. Either number alone is something a board can nod at and move past.

Two things, done visibly

The most common way to waste a climate survey is to take all 17 dimensions seriously at once. Seventeen improvement initiatives is no initiative at all, and by the time next year’s survey opens the crews will have noticed.

Crews judge a survey entirely by what visibly changes afterwards. Pick the two lowest dimensions. Do something concrete and specific about them. Tell the fleet what you found, what you decided, and what you did, including the parts you decided not to act on and why. Then measure again next year and show the movement. Response rates and honesty both follow from that loop, and both collapse without it. A survey that produces no visible change teaches the crew that answering honestly is a waste of a watch below, and that lesson is very hard to unteach.

The written organisation and the lived one will never be identical, and the gap is not a scandal; it is the normal condition of any company with ships at sea and people ashore. What matters is knowing how wide the gap is, where it runs, and which direction it moved since last year. The audit will keep telling you about the company on paper. This tells you about the other one.

If you want to see what your own fleet’s profile looks like, your first survey is free, and it covers the whole fleet: we set it up, you put the posters in the messrooms, and you get the full results at no cost, from the ranked dimensions to the four-quadrant profile and the safety climate section. In return, your anonymized aggregates join the industry benchmark the comparison is built on.