VA/NVA Taxonomy: Classifying Activities

A taxonomy organizes observations using operational definitions. VA may mean direct transformation; necessary NVA, support work; and delays, states whose causes need investigation. The classification is a study convention, not an individual-performance diagnosis or an instruction to eliminate every non-transforming activity.

Design and precision

Define the unit—person, machine or workstation—and period. Use mutually exclusive categories for each unit and rules for simultaneous activities. A processing machine and a waiting operator are simultaneous states of different resources. Include an unclassifiable state with a reason and review procedure; forcing a code does not improve the data.

A hypothetical example might find 18% searching, but no interval can be inferred without sample size and design. An approximate proportion plan uses n=ceil(z²p₀(1−p₀)/e²). With p₀=0.5, z=1.96 and e=0.03, this gives 1,068 independent observations. Repetition of the interval procedure targets 95% coverage; it is not a 95% probability statement about a fixed parameter in this particular realized interval. NIST explains the roles of parameter, precision and design.

Randomization does not ensure complete objectivity. Misclassification, missing coverage, reactivity and dependence remain possible. A larger sample does not cure those biases. The Hawthorne effect has no universal correction factor.

MECE is a design criterion, not a property established merely by using its name. Lean waste categories, MTM and duration measurement address related but different questions. Choose a method for the objective rather than declaring one classification universally superior.

Practical Architecture: The Three Categories and Their Operational Subcategories

The practical application of the taxonomy is structured in three levels, from executive aggregation to the analytical granularity needed to design concrete improvement actions. In WorkSamp, the taxonomy is configurable on a VA/NVA/IMP base, so each plant adjusts its categories to its own reality.

Value Added (VA): What Actually Transforms the Product

These are the activities that the customer explicitly pays for. They are the plant's reason for being. They are subdivided into:

VA-TC (Continuous Transformation): Continuous processes such as turning, plastic injection, or lamination.
VA-TD (Discrete Transformation): Unitary operations such as manual component assembly.
VA-PI (Information Processing): Actions that, based on data, physically modify the product (e.g., configuring a robot according to a program).
VA-VF (Integrated Functional Verification): In-line controls that adjust the process in real time, being part of value creation.

Non-Value Added Essential (NVA-E): The Current Necessary Evil

Activities that do not generate value but, with current technology and regulations, are unavoidable. The goal is not to eliminate them, but to minimize and optimize them. Their key subcategories include:

NVA-E-SG (Safety): Donning PPE, Lockout/Tagout (LOTO) procedures.
NVA-E-CL (Mandatory Quality): Final dimensional inspection required by the customer in the contract.
NVA-E-MN (Preventive Maintenance): Scheduled tool change, essential lubrication.
NVA-E-TR (Minimum Transport): Part movement between two adjacent machines in the current flow.
NVA-E-AD (Administrative): Mandatory traceability recording per regulations (e.g., aerospace, pharmaceutical industry).
NVA-E-CP (Process Change): Tooling adjustment between different batches.

Non-Value Added Avoidable (NVA-EV): The Golden Vein of Productivity

These categories identify candidates for investigation. Searching may serve a legitimate need, and conversation may be coordination or an agreed break. Do not assign individual fault or remove safety, quality, breaks or training merely because of a code. The following examples are a proposed classification requiring local boundaries and validation.

NVA-EV-BU (Searches): Searching for tools, documents, material, or the supervisor.
NVA-EV-ES (Waits): Waiting for authorization, material from a previous stopped machine, or instructions.
NVA-EV-TR (Unnecessary Transport): Long trips to central warehouses, unoptimized routes.
NVA-EV-RW (Rework): Reprocessing due to an avoidable error in a previous operation.
NVA-EV-MO (Movements): Excessive reaches, unnecessary body twists, forced postures.
NVA-EV-ST (Stock handling): Observable work moving or managing inventory; accumulated inventory is a separate process measure, not a person’s activity state.
NVA-EV-SP (Over-processing): Polishing not required by the customer, double verification due to process distrust.
NVA-EV-CO (Communication to classify): Observed communication whose operational purpose is not established; do not assume unproductiveness or investigate personal content.
NVA-EV-PI (Wait with unknown cause): Observed waiting whose cause needs checking; it does not imply unjustified worker conduct.

EV codes organize improvement hypotheses. A reading assigned to one does not demonstrate avoidability: establish necessity, cause and conditions before setting a reduction target.

Beyond Classification: Integration with Executive Metrics

The true power of this taxonomy emerges when integrated with high-level indicators, providing a holistic view of efficiency.

Direct work and OEE are not equivalent

Direct-work share needs a definition and explicit denominator. “Tool in hand” does not necessarily include all useful maintenance work. Safety inspection or planning may be essential even if excluded from that numerator.

OEE is not the product of VA/NVA state shares. It needs availability, performance and quality: Q=good units/total units, P=(ideal cycle time×total units)/run time, and A=run time/planned production time. The sampled proportion inspecting or reworking is not Q. See the OEE calculation.

Sampling may provide context on states and delays. Production and quality counts must be collected separately. Induly provides working-time records, not a substitute for those counts or a demonstrated automatic integration.

Diagnosis without a promised return

WorkSamp supports observation planning, recording and distribution analysis. Precision depends on study design. Assess improvement outcomes by checking which activities can safely change, whether the bottleneck releases usable capacity, and which costs actually change.

Implementation and Best Practices

For the taxonomy to be effective, some key principles should be followed:

Clear and agreed-upon definition: Before starting, the team of engineers and supervisors must align on the exact definition of each subcategory. Is a tool change an NVA-E-MN or an NVA-E-CP? These boundaries must be clear.
Observer training: The observer must be impartial and trained to classify consistently and quickly, without influencing operators' behavior.
Genuine randomness: Define appropriate selection probabilities and times; review routes to avoid synchronization with cycles. Randomization does not eliminate every bias.
Analysis and action: Data is useless without analysis. The value lies in the Pareto analysis of the main NVA-EV categories and in the implementation of countermeasures (5S for searches, line balancing for waits, etc.).

The taxonomy is useful when categories, units and resulting decisions are understandable and reviewable. Its quality depends on implementation and study limitations; it is not universally the “most objective” tool for every question.

Resources and Tools

To delve deeper into the practical application of these methodologies, we recommend exploring the following resources:

ASETEMYT Directory: Find providers, consultants, and specialized tools in industrial timekeeping and work study.
ASETEMYT Blog: Technical articles and case studies on productivity, methods, and times.
WorkSamp: Application for planning, recording and analyzing work-sampling observations.
Cronometras: Specialized software for conducting time and motion studies in an agile and precise manner.
Induly: Working-Time and Industrial Timekeeping platform (working hours, shifts and clock-ins).
Want to add your tool or service to the directory? Add your company here.

ProdCont: Industrial productivity consulting; comprehensive operational audits (not just work sampling) for plants with several open fronts.

Statistical source

NIST: selecting sample sizes and design assumptions.