A job that takes 45 seconds per part, runs on two shifts and regularly creates a queue at the end of the line is a stronger automation candidate than a task that takes ten minutes but only occurs twice a week. For a production manager, the useful question is: when is automation profitable? The answer comes from the operating data around the task, not from the robot catalogue.
A collaborative robot can improve output, consistency and working conditions, but only when the full cell is designed around a stable process. The right starting point is therefore not payload or reach. It is the cost of the current process, the volume that must be handled, and the constraints that prevent the team from producing more profitably today.
When is automation profitable? Start with the economic case
Automation is profitable when the annual financial gain exceeds the annualised cost of the installed solution within a payback period your business accepts. That sounds obvious, but many calculations fail because they compare a robot's purchase price only with an operator's wage.
The real comparison is between two production systems. The manual system includes direct labour, supervision, rework, scrap, stoppages, variable cycle times, injuries or fatigue risk, and the opportunity cost of assigning skilled people to repetitive work. The automated system includes the cobot, end-of-arm tooling, guarding or safety measures where required, fixtures, integration, validation, maintenance, training and expected downtime.
A practical calculation begins with annual productive hours. Multiply the time removed from a manual operation by the loaded hourly labour cost, then add measurable savings from reduced defects, less material waste and higher machine utilisation. Against this, include the full installed cost and realistic ongoing costs. If the cell frees an operator to perform higher-value work rather than removing a role, count the value of that newly available capacity, not an assumed headcount reduction.
For example, a screwdriving operation may save only one operator's partial attention. Yet it can still justify itself if torque traceability avoids expensive field failures, the robot keeps a machining centre supplied during breaks, and the former operator can solve exceptions across several cells. Profitability is often created by combined gains rather than one dramatic labour saving.
Use a payback target, but do not stop there
Many manufacturers look for a payback period of 12 to 24 months. This can be a sensible filter, particularly where product demand or process stability is uncertain. A longer payback may still be justified for safety-critical quality, documented process control or capacity that supports a confirmed customer programme.
Payback alone does not show the whole picture. A cell with a quick payback but poor flexibility may become a stranded asset after a product change. Assess the likely useful life of the workstation, expected product mix, maintenance requirements and residual ability to redeploy the cobot to another operation. Collaborative automation is particularly valuable where redeployment is realistic rather than theoretical.
High-volume repetition is helpful, not the only condition
Repeatability is the foundation of profitable automation. The robot needs a consistent presentation of parts, a defined cycle, reliable gripper engagement and clear rules for what happens when something is wrong. If every tenth part arrives in a different orientation or requires judgement based on cosmetic variation, solve that variation first or include sensing and recovery in the scope.
High volume improves the case because setup costs are spread across more parts. However, low-to-medium-volume production can also be attractive when changeovers are short and programming is accessible to the production team. A cobot that can be reassigned between tending, packing and assembly tasks can earn productive hours across the week rather than waiting for one large batch.
The most promising applications commonly share several characteristics:
The task is repetitive and physically demanding, such as machine tending, palletising, welding or screwdriving.
Cycle times are stable enough to define a reliable sequence and a realistic throughput target.
Quality depends on consistent force, position, torque or timing.
Labour availability limits output, shift coverage or machine utilisation.
Parts and fixtures can be presented consistently without excessive manual preparation.
If only one of these conditions exists, the business case may be weak. A highly repetitive task with no available production hours, for instance, can be difficult to justify. Equally, an ergonomically poor task may deserve automation even where the direct financial return is modest, but that decision should be made openly rather than dressed up as a rapid payback project.
Calculate capacity, quality and utilisation properly
A common error is to use the robot's theoretical cycle time as the expected production rate. Real throughput includes loading, unloading, gripper actions, machine handshakes, inspection, part replenishment, planned changeovers and recovery from faults. The relevant measure is good parts per hour over a representative shift.
Consider machine tending. If a CNC machine waits three minutes between jobs for an operator, reducing that waiting time can create capacity without buying another machine. But if the robot takes longer than the operator to load a component, the value must come from unattended running, extended shifts or improved consistency. The calculation changes with the production schedule.
Quality deserves the same discipline. A claimed reduction in defects should be linked to actual scrap cost, rework time, customer returns or inspection effort. In screwdriving, recorded torque and angle data may identify a process drift before defective assemblies leave the factory. In dispensing, controlled paths and repeatable speed can reduce material use. These are meaningful savings, but they should be quantified from your own production records.
Also separate capacity from demand. Producing 20 per cent more parts is profitable only if those parts can be sold, replace overtime, reduce subcontracting, or protect delivery performance. When demand is variable, flexible deployment and fast changeover may be worth more than maximum speed.
The hidden cost is usually process variation
The robot is rarely the difficult element. Uncontrolled inputs are. Parts stacked inconsistently, worn fixtures, changing materials, undocumented operator adjustments and unreliable upstream equipment all consume engineering time and erode availability.
Before approving a project, observe the operation across shifts. Record actual cycle times, minor stops, rejected parts, rework routes and the different ways experienced operators deal with exceptions. This work often reveals improvements that should be made before automation. It may also show that a relatively simple fixture, sensor or end effector will remove the source of variation.
A feasibility test should prove the critical uncertainties, not merely demonstrate that a robot can move between two points. Test the real parts where possible. Check reach, payload including the gripper, required accuracy, cable routing, surface condition, process forces, access for replenishment and recovery after a fault. For welding, assembly and insertion, force behaviour and part tolerances may matter more than headline repeatability.
FAIRINO Europe approaches this stage through technical feasibility assessment, workstation design and live testing because an honest result is valuable even when the outcome is that the application should wait. A dedicated machine, a manual improvement or a different process sequence can sometimes be the better answer.
Choose the right level of automation
A collaborative robot is not automatically the most profitable choice simply because people work nearby. Cobots are well suited to flexible cells, moderate speeds, shorter runs, frequent product changes and operations where direct interaction or a compact footprint is useful. Their accessible programming can reduce dependence on specialist resources, especially when internal teams need to adjust proven routines.
A fixed, high-speed process with long, stable production runs may justify dedicated automation instead. Conversely, a complex task requiring human judgement, delicate visual assessment or frequent unstructured handling may remain best performed by an operator, perhaps supported by a fixture or digital work instruction.
Safety design must be part of this decision. Collaborative operation does not remove the need for a risk assessment. Tool geometry, sharp edges, payload, speed, pinch points and the wider workstation can require safeguarding measures. Budgeting safety and operator access from the outset prevents late changes that undermine the expected return.
Build an implementation plan that protects the return
The most profitable cells are designed for ordinary production, not ideal demonstrations. Define who refills materials, who responds to alarms, how quality is checked, where rejected parts go and who can restart the cell safely. Train operators and maintenance staff before commissioning, then give them ownership of the standard work.
Commissioning should include an agreed acceptance test: target cycle time, good-part rate, traceability requirements, uptime expectations and changeover procedure. Track performance during the first weeks and distinguish between robot downtime, upstream shortages and process-related stops. Without this data, a sound investment can be wrongly judged by a poorly prepared launch.
A good next step is to select one constrained operation and collect a week of real production data. Bring the parts, cycle information and quality records to a feasibility discussion. The result should be a clear decision: automate now, improve the process first, or keep the task manual until demand and repetition justify the investment.