Planning your first automation investment

We sell machines, so you might expect this article to say “automate everything.” It will not. We have watched factories buy robots that became coat stands, and we have watched a single well-chosen machine transform a business. The difference is never the machine. It is the selection process that came before the purchase order.

Automate the boring, stable, measurable thing

The ideal first automation candidate scores high on all four:

1. Repetition — thousands of identical cycles per shift, not dozens of varied ones.
2. Stability — the product and process will not be redesigned next year. Automating a moving target burns money.
3. Measurability — you can state today's cost per part, defect rate and throughput. If you cannot measure the before, you will never prove the after.
4. Pain — the station where quality escapes happen, where recruitment is hardest, or where one absence stops the line.

In hardware manufacturing this points to the same shortlist again and again: screw and fitting insertion, riveting, packing and cartoning, and repetitive inspection. In furniture factories it points to edge-band finishing, drilling and dowel insertion, and panel handling between machines — the lifting-and-turning work nobody applies for anymore.

Payback maths, done honestly

The brochure calculation is machine price divided by saved wages. The honest calculation adds what the brochure omits:

• tooling, fixtures and spares — commonly 15–25% on top;
• installation, commissioning and lost output during ramp-up;
• an operator who becomes a machine minder, not zero labour;
• maintenance skills you may need to hire or train.

And it credits what the brochure also omits: scrap that stops happening, claims that stop arriving, capacity released without extending the building, and output that no longer varies with the night shift's mood. In our experience the quality-side benefits routinely match the labour savings — but only for processes chosen by the four tests above.

A machine that saves two operators but needs a better technician you cannot hire has saved you nothing.

The mistakes that stall projects

Automating chaos. If the manual process has no stable standard — parts arrive mixed, tolerances drift, operators each have their own method — automation will faithfully reproduce the chaos at higher speed. Standardise first, automate second.

Ignoring the material flow. A fast cell fed by a slow forklift is a slow cell. Layout and buffering decide real throughput more often than cycle time does.

Buying capability instead of capacity. Six-axis robots are seductive. Very often a cam-driven insertion unit at a fifth of the price does the actual job faster and never needs a programmer.

No acceptance criteria. Agree measurable output, quality and uptime figures in the purchase contract, demonstrated on your parts before shipment and again after commissioning. A vendor who resists that clause is telling you something useful.

Start where we can help

Klinq designs hinge and slide production lines, and automation cells for furniture hardware assembly. If you send us your target products and volumes, we will tell you which stations justify automation now, which should wait — and we will put output guarantees in writing.