Work Sampling in the Naval Sector: How to Audit Productivity in Shipyards with Multiple Concurrent Teams

Introduction: The Shipyard Is the Most Complex Environment for a Sampling Study

Most of the literature on Work Sampling has been developed on "easy" industrial environments to observe: a manufacturing plant with stable lines, an assembly chain with repetitive cycle, a call center with equivalent workstations. Work Sampling, as formulated by Tippett in 1935 and refined by the British Standards Institution and the ILO in the following decades, assumes that the workstation or area to be observed is reasonably homogeneous and stationary: the operator is doing "the same" during the observation period, the activity categories are stable, and the sample variability is due to random sampling, not to structural process changes.

The shipyard breaks all those assumptions. A 300,000 deadweight ton tanker (VLCC) or a 5,000-passenger cruise ship is built with thousands of operators working simultaneously on a single physical unit that moves and transforms. The work structure changes every week. A hull block that was in welding yesterday is in grit blasting today. A welding team that worked on the stern on Monday is on the bow on Friday. The trades — fitters, welders, pipefitters, electricians, painters, boilermakers, insulation fitters, instrumentation technicians — cross in the same physical space with dependencies, waits, and planning conflicts that do not exist in a conventional production line.

For all these reasons, Work Sampling in the naval sector requires serious methodological adaptations compared to the classic scheme. Applying the standard formula N = 4p(1-p)/e² without more, without stratifying by structural block, without distinguishing trades, without controlling the concurrency effect, produces results that seem rigorous and are not. This article explains why, what to do about it, and how to use sampling to reduce naval construction costs — where every percentage point of unproductive translates into hundreds of thousands of euros of budget deviation in a single ship.

1. The Four Peculiarities of the Shipyard That Invalidate "Standard" Sampling

1.1. Unique Product, Not Series

In an automotive plant, 1,000 units per day of the same model are produced. A shipyard produces one unit per year (or less) of a product that has never been built before. This has two important statistical consequences:

There is no replicable history. You cannot compare the welding time of a block of ship 100 with that of ship 101, because ship 101 has another block type, in another position, with another accessibility.
Variability between "pieces" is enormous. A structural block can vary in welding time by a factor of 3x between a flat block and a block with double curvature, or between a small block (8 tons) and a large block (250 tons). The sampling variance inflates.

1.2. Multitude of Trades Working in Concurrency

In a merchant ship construction shipyard, at the peak of the work you may have 20-40 different trades working on the same block or section. This creates an impossible observation problem for a single observer: you cannot instantaneously classify what each of the 40 operators present in a 200 m² block is doing. Stratification is necessary.

1.3. Outdoor Work, Environmental Conditions

Naval construction is an almost always outdoor activity. Rain, wind, heat, cold, humidity modify productivity and safety. The observer must record the environmental condition on each observation, because the "valid" categories on a rainy day are different from those on a dry day. A rainy day can triple the proportion of "weather wait" and that must be isolated from the rest.

1.4. Hierarchical Work Structure: Block → Section → Module → System

The shipyard does not organize work by "cycle times", but by structural blocks (steel pieces that are pre-assembled in workshop), which are joined into sections (block assemblies), which are mounted into modules (complete functional blocks: galley, engine room, bridge), and which are integrated into systems (electricity, plumbing, air conditioning, instrumentation). Each level has Productivity, Standard Times, and VA/NVA categories that are different. Sampling must respect this hierarchy or it will lose traceability with the shipyard's planning system.

2. Methodological Adaptations: How to Make Work Sampling Applicable to a Shipyard

2.1. Stratification by Structural Block and Trade

The first step — non-negotiable — is to stratify the population to be observed. If the shipyard has 8,000 operators and 1,500 blocks in the ship, you cannot do a "shipyard" sampling. A sampling by block and by trade is done, with N calculated for each cell of the double-entry table. This implies more total observations, but each stratum is internally homogeneous, so the N per stratum is reduced.

Example of typical stratification for a Suezmax tanker in block phase:

Trade	Small block (<20 t)	Medium block (20-100 t)	Large block (>100 t)	Subtotal
Welding	220	350	480	1,050
Fitting	180	290	380	850
Piping	140	220	310	670
Painting/Grit Blasting	100	160	240	500
Electricity	80	120	180	380
Others	60	90	150	300
Total per block	780	1,230	1,740	3,750

With this table, N is calculated for each cell according to the formula N = 4p(1-p)/e². For p=0.5 and e=0.05, each cell requires 400 observations. For p=0.3, 336. The total study figure (sum of cells) is 3,000-4,000 observations, distributed over 4-6 weeks to cover all shifts and environmental conditions.

2.2. Randomness Controlled by Schedule and Shift

The randomness of observations is the key to the statistical validity of sampling. But in a shipyard it is not enough to choose random instants: they must be distributed between the three shifts (morning, afternoon, night), working and non-working days, and the different phases of the work cycle. If you only observe morning shift, you are measuring a fictitious shipyard: the night has its own distribution (more maintenance tasks, less concurrency, more logistical waits).

A weekly observation plan is generated with a random seed to avoid observer bias. Applications like WorkSamp automate this planning, generating random observation routes stratified by block/trade/shift and registering with time stamp and geolocation each observation. This is critical in shipyards where audit control is required for ISO 9001 certifications and for the traceability of man-hours billed to the shipowner in turnkey contracts.

2.3. VA/NVA Taxonomy Adapted to the Shipyard

The classic VA (Value Added) / NVA (Non Value Added) taxonomy requires naval adaptation. A standard proposal, based on the recommendations of the NSRP (National Shipbuilding Research Program) of the United States and the guides of SEA Europe (Shipyards' & Maritime Equipment Association of Europe):

VA activities (generate value for the shipowner):

Structural block welding (whose result remains in the ship)
Block fitting and assembly
Final piping installation
Equipment electrical wiring
Main equipment assembly (engines, generators, pumps, critical valves)
Functional tests (FAT, SAT, sea trials)

NNVA activities (necessary but not generating final value):

Block movement between workshops
Material transport to the block foot
Scaffolding assembly and disassembly
Tool and fixture preparation
Drawing and/or 3D model reading
Weld validation (NDT, non-destructive testing)

NVA activities (do not generate value, must be eliminated):

Waits due to interferences with other trades
Weather waits (rain, wind, waves)
Waits due to lack of material or spare parts
Re-work due to welding, fitting, or installation errors
Unnecessary movements (looking for tools, materials, drawings)
Unpaid unregistered breaks

In practice, reference shipyards achieve VA distributions of 35-45%, NNVA 30-35%, and NVA 20-30%. An average shipyard moves around 25-30% VA, 30-35% NNVA, and 35-40% NVA. The difference between an efficient shipyard and an average one translates into 3-6 months of advance in ship delivery and millions of euros in savings in labor and consumables.

2.4. Sampling Stratified by Construction Phase

The shipyard is not measured the same in the steel preparation phase (cutting, rolling, profiling) as in the block fitting phase, as in the grit blasting and painting phase, as in the launch and outfitting phase, as in the sea trials phase. Each phase has its own time distribution, its own labor intensity, and its own complexity. Sampling must be designed by phase, with observations concentrated in the phase that most affects the total project cost: typically the block fitting and welding phase (40-50% of total man-hours) and the outfitting and assembly on the slipway phase (30-40%).

2.5. The N Formula Still Holds, but the "p" Becomes Dynamic

The essential formula of sampling N = 4p(1-p)/e² remains intact, but the value of p (expected proportion of the activity we want to measure) must be estimated per cell through a prior pilot study. In naval, a pilot study of 200-300 observations per trade/block is essential to estimate p with sufficient precision. Without a pilot, the N is over-dimensioned (and money is wasted) or under-dimensioned (and biased results are obtained).

3. Illustrative Case: Audit of a Suezmax Tanker Hull Block

Note: the data, figures, and names in this section are an anonymized illustrative example, built from the usual casuistry of medium-sized shipyards in southern Europe. Any coincidence with real projects is casual.

3.1. Context

Medium-sized shipyard (capacity: 8-12 ships/year), building a Suezmax tanker of 158,000 DWT. Phase: block fitting and welding in workshop. Selected block: double bottom block of the central tank, 85 tons, accessibility limited by its position in the workshop. 14 welders, 4 fitters, 2 pipefitters assigned to the block in the study period.

3.2. Study Design

Period: 21 working days, distributed over 5 weeks to cover all lunar and weather phases of the season.
Total observations: 2,730 (195/working day, 78/shift).
Stratification: 4 trades × 3 block types (small, medium, large) = 12 cells. The example block falls into the "welding / medium block" cell.
Target confidence level: 95%, acceptable error ±4% for the main categories.
Categories observed: 9 (VA welding, VA fitting, NNVA preparation, NNVA movement, NVA interference wait, NVA weather wait, NVA material lack, NVA rework, NVA unpaid break).

3.3. Results

For the "welding / medium block" cell:

Category	% observed	95% CI
VA welding	48%	[44%, 52%]
NNVA preparation (cleaning, fixtures)	16%	[13%, 19%]
NNVA movement (material transfer)	8%	[6%, 10%]
NVA interference wait (other trades)	9%	[7%, 11%]
NVA weather wait	4%	[3%, 5%]
NVA material lack	5%	[3%, 7%]
NVA rework	6%	[4%, 8%]
NVA unpaid break	4%	[3%, 5%]

Interpretation: the block is being built with 48% real value added. Adding NNVA (24%), the block consumes 72% of the time on activities that ultimately provide value or are necessary. 28% of the time is lost in avoidable unproductive. On the 1,200 man-hours planned for this block, that is 336 wasted man-hours, which at an average naval industrial cost of €38-45/h, translates into €12,800-15,100 of avoidable cost in a single block. Multiplied by the 40-60 blocks of a Suezmax tanker, the avoidable cost of a single ship can be in the range of €500,000 to €900,000.

3.4. Derived Action Plan

Attack "interference wait" (9%): the block is waiting for other trades 9% of the time. Coordinate trade planning with the shipyard's 4D planning system. WorkSamp can help prioritize bottlenecks per block/trade cell.
Reduce "rework" (6%): 6% of hours go into correcting defective welds. Root cause analysis with NDT (non-destructive testing) and continuous welder training. Here the integration with a weld quality control system with digital traceability is key.
Optimize "preparation" (16%): a high NNVA can indicate that fixtures, electrodes, or drawings are not at the site when the welder needs them. Apply 5S at the block foot and study the workshop layout.
Establish a quarterly audit cycle: apply sampling every 3 months to detect regressions and validate improvements.

4. The Complete Audit Ecosystem in a Shipyard

Work sampling in the naval sector is not done alone. The usual — and most efficient — thing is to combine several specialized tools, each with a specific role in the continuous audit flow of a shipyard. The three key pieces of the ProdCont ecosystem are:

4.1. Cronometras — Time Studies by Stopwatch

The Cronometras application is the reference tool for time and motion studies with professional stopwatch. In a shipyard, it is used when you need to establish or update the standard time of a specific operation: for example, timing 30 cycles of fillet weld welding in horizontal position to obtain the NT (normal time) and then apply ILO allowances. The particularity of naval is that the variability of the operation is high (depends on position, accessibility, joint type, electrode diameter, current, etc.), so more timing cycles are needed than in a standard factory environment. Cronometras automatically calculates the remaining takes to reach a 95% confidence level according to the statistical method, which reduces the number of unnecessary timings. Its Multi Crono module (up to 6 simultaneous stopwatches) is especially useful in shipyards where you want to time several welders in parallel on the same block.

4.2. Induly — Time Control and Presence Registry

Induly provides the industrial time control on which any bonus, any attendance KPI, and any audit of billable man-hours is calculated. In a shipyard, where man-hours are the main cost variable (60-70% of the cost of a ship), the traceability of real hours worked is critical. Induly allows recording entries and exits by trade, block, shift, with plant kiosk clock-in, geolocation, and task validation. This gives you the denominator (how many hours have been worked) that work sampling complements with the qualitative numerator (in what those hours have been invested). Without Induly, work sampling operates on aggregated or paper presence data, which introduces 5-15% errors in the final figures.

4.3. WorkSamp — Specialized Work Sampling

WorkSamp is the application specialized in Work Sampling, designed specifically to execute work sampling studies with the depth required by a shipyard. It allows:

Stratify observations by block, trade, and shift.
Generate random observation routes with optimal frequency (based on Poisson statistical independence analysis, not arbitrary rules).
Register offline observations from tablet or mobile, with time stamp and geolocation per block.
Calculate 95% and 99% confidence intervals per cell.
Generate executive reports with the VA/NNVA/NVA distribution per cell, with comparison between blocks, trades, and construction phases.
Export to Excel and PDF to integrate with the shipyard's planning system (Primavera P6, NaviWorks, or similar).

WorkSamp is the only tool on the Spanish-speaking market designed specifically to execute stratified sampling at the scale required by a shipyard (3,000-10,000 observations), with the traceability needed for an external audit or presentation to the shipowner.

4.4. ASETEMYT — Naval Consultants Directory

The ASETEMYT directory includes consultants and specialized companies in naval productivity improvement. If your shipyard needs external support to design a large-scale sampling study, validate results, or train your team in the methodology, there you will find professionals with demonstrated experience in the sector.

4.5. ProdCont — The Strategic Vision of the Ecosystem

ProdCont is the umbrella brand on which the three previous tools depend. If you want to understand how Cronometras, Induly, and WorkSamp integrate into a continuous improvement strategy for the whole shipyard (not just a block or a ship), the ProdCont website explains the global methodology.

5. The Pilot Test: Why Start with One Block, One Trade

A frequent mistake when introducing Work Sampling in a shipyard is trying to do it "big" from day one: 8,000 observations, 30 trades, 5 phases. This usually ends in a study that takes 6 months, costs a fortune, and produces results that nobody knows how to interpret.

The correct strategy is start small, learn, scale. The usual plan in a medium shipyard:

Month 1: pilot study in a single structural block of a single trade (welding, for example). 400-600 observations. Team of 2-3 observers. Result: learn the mechanics, validate the taxonomy, calibrate the CI.
Months 2-3: extend to 3-4 representative blocks of the fitting phase. 1,500-2,000 observations. Validate stratification.
Months 4-6: extend to all blocks of the fitting phase. 5,000-8,000 observations. Executive report for management and for the shipowner.
Month 7 onwards: quarterly audit with 1,000-1,500 observations. Follow-up sampling.

This plan is executed entirely with WorkSamp, from pilot planning to quarterly audit. The typical learning curve is 2-3 weeks for a team of 2-3 people with basic training in methods engineering.

6. Conclusion: Work Sampling as a Naval Management Tool

Work Sampling, with almost 90 years of history since its original formulation, remains one of the most powerful tools to understand and improve industrial productivity. But applied to the naval sector without the necessary methodological adaptations (stratification, specific taxonomy, phase control, prior pilot), it produces results that seem rigorous and are not.

The good news is that, with the appropriate adaptations, naval sampling works extraordinarily well and provides measurable savings of hundreds of thousands of euros per ship. The key lies in:

Stratify by block and by trade, not treat the shipyard as a homogeneous whole.
Adapt the VA/NVA taxonomy to real naval vocabulary.
Start with a small pilot before scaling.
Combine sampling with professional time study (with Cronometras) and rigorous time control (with Induly) to have the complete picture.
Audit every 3 months to maintain continuous improvement.

If your shipyard is embarking on a new project or wants to diagnose the productivity of one in progress, the most direct path is:

Request a WorkSamp demo and learn about the tool with a free 400-observation pilot in a single block.
Review the VA/NVA taxonomy proposed in this article and adapt it to your shipyard.
Measure, decide, measure again. Naval continuous improvement is not a declaration of intentions: it is a quarterly cycle of statistical audit.

The ProdCont ecosystem — Cronometras, WorkSamp, and Induly — is what shipyards and engineering companies in the region use to address productivity with statistical rigor. If your shipyard is taking the step toward industry 4.0 without neglecting the rigor of classic work study, this is the path.

Do you want to start with a Work Sampling pilot in your shipyard? Request a free WorkSamp demo and we will help you design the first study in less than 2 weeks.