Low-Volume, High-Mix Manufacturing: The Automation Problem Nobody Solved

For forty years, industrial robot automation has worked on one premise: run the same task, on the same part, at high volume, long enough to justify the programming cost.

That premise fits automotive tier-1 suppliers. It fits high-volume consumer electronics assembly. It fits continuous process industries.

It does not fit the majority of manufacturing businesses.

What Is Low-Volume, High-Mix Manufacturing?

Low-volume, high-mix (LVHM) — sometimes called high-mix, low-volume (HMLV) — describes manufacturers who make a wide variety of different products, each in relatively small quantities. Common sectors include:

• Job shops and contract manufacturers — custom parts to customer drawings, batch sizes from 1 to 1,000
• Aerospace components — many different part numbers, strict traceability, low volumes per part
• Medical device manufacturing — diverse SKU range, tight tolerances, controlled processes
• Industrial machinery builders — made-to-order equipment, each unit somewhat unique
• Specialty automotive — classic vehicle restoration, racing, specialty vehicle builds

These manufacturers share a characteristic: their production mix changes faster than traditional automation can follow.

Why Traditional Automation Fails LVHM

The maths of traditional robot automation are straightforward:

• Cost of automation: €15,000–€50,000 per task (programming + integration)
• Minimum batch size to break even: depends on cycle time and part value, but often tens of thousands of units

For an LVHM manufacturer running 500-unit batches across 30 different part numbers, the economics simply do not work. By the time a robot is programmed and commissioned for part A, the order for part A is finished and part B needs the machine.

The Changeover Problem

Even when automation is initially justified, the changeover cost kills the case. Every time the product mix changes:

• New task = new integrator call
• New integrator call = new invoice (€5,000–15,000)
• New invoice = new wait (3–5 weeks)
• New wait = the line works around the gap

After the first or second changeover cycle, most LVHM manufacturers conclude that robot automation is not for them. The robot sits doing one task while a dozen other tasks wait for integration that never comes.

The Knowledge Problem

When an integrator programs your robot, the knowledge of why the program works goes with them. Adjustment A at line 47 prevents the arm from colliding with the part during a specific orientation — but that is not written anywhere. When you need to modify the task, you need the integrator back, because only they know what their choices mean.

In an HVLM environment, where tasks change constantly, this knowledge concentration is a structural liability.

What Is Actually Required for LVHM Automation

A robot that can serve LVHM manufacturing needs fundamentally different properties than one that serves high-volume, low-mix:

1. Fast task changeover. Ideally same-day. If it takes weeks to reprogram, it cannot serve batches that complete in days or weeks.

2. Worker-owned programming. The person doing the task needs to be able to teach the robot the task. Depending on an external specialist for every changeover is incompatible with LVHM economics.

3. Low per-task programming cost. If each new task costs €5,000–15,000 in programming, and you need 20 new tasks per year, that is €100,000–300,000 in annual programming spend on top of the hardware. That is not a business model that works.

4. Flexible end-effectors. LVHM manufacturers handle many part geometries. Grippers need to handle variation — different part widths, weights, and surface finishes — without requiring a custom end-effector per part family.

5. No specialist dependency. When the specialist is unavailable (sick, busy, expensive), the robot cannot be changed. LVHM requires that the factory team can manage the robot independently.

What Has Changed

Until recently, meeting all five of these requirements simultaneously was impossible. Robot programming required specialists. Specialist dependency was the price of automation.

Two developments have changed this:

1. Collaborative Robots (Cobots)

The introduction of force-limited collaborative robots from Universal Robots (2008) and subsequent manufacturers removed the hard safety barriers that made traditional industrial robots inaccessible to factory workers. Workers can now be in the robot's workspace — the robot will stop if it contacts a person.

This removed the safety engineering overhead that made small-batch automation impractical. But cobots still required specialist programming via teach pendant or code.

2. No-Code and Low-Code Programming

The second development is AI-powered programming — platforms that allow workers to teach robots by demonstration rather than by programming. A worker demonstrates the task using a phone camera and voice narration. The system converts the demonstration into a deployable robot program.

This breaks the specialist dependency. The factory worker who does the manual task can now also teach the robot to do it.

The Business Case for LVHM Automation Today

With no-code programming, the economics of LVHM automation change substantially:

For a job shop running 30 different tasks per year across 3 robots, the difference in total programming cost over three years is measured in hundreds of thousands of euros.

Practical Guidance for LVHM Manufacturers Evaluating Automation

Start with the right task

Not every task is a good first candidate. The best first task for LVHM:

• High labour intensity relative to value added (the worker spends most of their time on the physical motion, not on judgment calls)
• Consistent enough to automate (parts that always look the same, arrive in the same orientation)
• High enough annual volume that the robot will actually pay for itself

Test changeover, not just setup

When evaluating a no-code platform, ask specifically: "How long does it take to change from task A to task B?" If the answer involves calling anyone outside your facility, the platform is not suited for LVHM.

Confirm multi-step support

LVHM tasks are rarely simple. Machine tending, assembly, and quality routing all involve multi-step sequences with conditional logic. Confirm the platform handles this before committing.

Plan the gripper strategy

With many part geometries, a single gripper will not serve all tasks. Develop a gripper strategy: a small inventory of grippers covering your part families, with a documented changeover procedure for the robot operator.

The Bottom Line

Low-volume, high-mix manufacturing has been left behind by traditional robot automation because the economics never worked. Programming cost, changeover cost, and specialist dependency together made automation inaccessible for anyone running batches under a few thousand units.

No-code programming changes the fundamental economics. The question is no longer "can we afford to automate?" — it is "which tasks should we automate first, and in what order?"

That is a much better question to be asking.

See also:

• Why Robot Tasks Still Are Not Automated — and What Changes That
• A Practical Guide to Robot Retasking Without an Integrator
• What to Ask Before You Buy a Robot Automation Solution