2026 Operations Study

For years industrial networks were designed for the cost optimum: single sourcing at lowest unit cost, minimal inventory, high utilization, long transport routes at low freight rates. That design was rational as long as demand and supply were predictable.

Under volatility the calculation flips. The cost of a supply interruption — expedited freight, replanning, contractual penalties, lost orders — regularly exceeds the unit cost saved by a multiple. Resilience is therefore not a soft add-on requirement but a design decision with its own business case.

In practice this means second sources for critical parts, a deliberate inventory strategy by criticality rather than across-the-board reduction, and transparency on the real supplier structure down to tier two.

Of all operational levers, working capital acts on liquidity fastest. The reason is structural: inventory, receivables and payables are already committed capital that can be released — without investment, without lead time, without employment law timelines.

Three entry points carry most of the impact: inventory coverage by item criticality rather than blanket targets, disciplined receivables management with clear escalation stages, and payment term governance towards suppliers.

Durability is what matters. Cutting inventory just before the reporting date is not an improvement but a shift. Proof must be provided on averages, not on the reporting date.

The move from trial to productive use is highly uneven across industrial operations. In three areas it has largely happened: production planning and demand forecasting, quality assurance with visual inspection, and technical service with aftersales documentation.

These areas share three properties: sufficient and structured data, clearly measurable benefit, and limited change required to existing processes. Where one of those conditions is missing, applications stay in pilot regardless of model quality.

A simple test follows for prioritization: is the data there, is the benefit measurable, is the process changeable? Three yeses mean scaling. One no means preparatory work, not abandonment.

Many automation decisions of recent years rest on assumptions from a period of low capital cost. In that constellation almost any substitution of labour by capital paid back.

With changed funding costs the threshold moves. At the same time wage pressure pushes the other way. The result is not a blanket verdict but a need for case-by-case calculation: equipment with high utilization and a stable product mix still pays back, while high-variance, low-frequency applications often no longer do.

The practical consequence is uncomfortable but important: an automation project approved three years ago and not yet implemented should be recalculated before implementation — not merely rebudgeted.

The recurring observation in operations programs is that the difference between plants with comparable equipment and product structure rarely lies in technology. It lies in the leadership routine.

Standard work, a functioning loop of deviation, root cause and countermeasure, and shopfloor management that makes deviations visible rather than administering them — these elements explain more of the performance gap than any single investment.

The sequence follows: stabilize the leadership routine first, then add technology. Putting technology on an unstable process automates the variation along with it.