Effective Routing and Managing Production Scheduling

The problem with a static plan

A high-mix, low-volume job shop has a different scheduling problem from a high-volume production line. On the production line, the plan is set, the sequence repeats, and disruptions are the exception. In a job shop, disruption is the norm. The job mix changes daily. Machines that were free yesterday are loaded today. A rush order arrives at 2pm and the afternoon schedule is already wrong.

Jon Scheele and Sim Thiam Soon from Qmani explored how manufacturers can build a scheduling system that stays useful even when the plan changes — drawing on SQP Engineering’s implementation of infoboard as a practical case study.

The utilisation vs. flexibility dilemma

The central tension in high-mix scheduling is between two things you want simultaneously: high machine utilisation (keeping equipment busy and revenue flowing) and operational flexibility (being able to absorb a rush order without disrupting everything else).

These pull in opposite directions. Maximum utilisation means scheduling every available minute, which leaves no room to move when something changes. Maximum flexibility means holding capacity in reserve, which looks like idle time on the floor and inefficiency on the books.

The session mapped this as a 2x2: low utilisation with high flexibility (responsive but wasteful), high utilisation with low flexibility (efficient but brittle), low utilisation with low flexibility (the worst of both — this is what happens when a business is overwhelmed but still can’t respond to customers), and the target zone: high utilisation with high flexibility. Getting there requires deliberate choices about how you structure work, not just better scheduling software.

Four approaches that move you toward the target zone

Family-based scheduling

Rather than treating each job as completely unique, manufacturers can group similar jobs that share common tooling, fixtures, or processes. At SQP Engineering, components that differ only in length but use identical tooling are scheduled consecutively to minimise changeover time. The result: less setup waste, more machine time on productive work.

Family-based scheduling also improves cost accuracy. When similar jobs run together, actual cycle times get captured in a consistent context — which means the estimates used for future quoting get better over time.

Flexible resource management

The traditional “one person, one machine” model creates scheduling rigidity. If your CNC operator is the only person who can run the CNC, then every absence or rush job that hits the CNC creates a crisis. Cross-training operators to work across machines gives the scheduler options — and means the bottleneck can sometimes shift to wherever the work is, rather than being fixed in one place.

Planned flexibility windows

Instead of scheduling every available minute of machine time, leave intentional gaps. These aren’t idle time — they’re capacity insurance. A rush order that arrives at 2pm doesn’t need to destroy the rest of the week’s schedule if there’s a window available tomorrow morning.

Setup time reduction

The smaller your batch sizes, the more setups you run — which means setup time directly constrains your ability to be flexible. Investing in quick-change tooling, standardised fixtures, and pre-staged materials reduces the cost of each changeover, which in turn reduces the pressure to run large batches just to amortise setup time.

Managing bottlenecks: Theory of Constraints in practice

The Theory of Constraints, developed by Eliyahu Goldratt, gives a framework for identifying and exploiting the resource that limits your throughput. In any production system, one resource — machine, person, or process — constrains the output of the whole. Improving non-bottleneck resources doesn’t increase output; it just creates more WIP in front of the bottleneck.

The ToC approach has four steps. Identify your constraint — the machine or person that jobs queue in front of most often. Exploit it — make sure it is never idle for want of work, and never produces defects that consume its time. Subordinate everything else — other work centres should pace themselves to feed the constraint, not optimise their own utilisation. Then, if output is still insufficient, elevate — add capacity at the constraint specifically.

In a high-mix shop, the bottleneck can shift depending on the job mix in any given week. A machine that’s comfortably loaded on a standard week can become the constraint when a particular type of job comes in. Sim Thiam Soon discussed how real-time visibility — knowing what’s queued and what’s running — is what makes the constraint visible before it becomes a crisis.

SQP Engineering: from paper to digital

Production Manager Rastko Petrovic shared SQP’s journey from paper-based scheduling to infoboard. The most immediate benefit wasn’t smarter scheduling — it was visibility. Being able to see machine loading, material status, and job progress on a single screen changed the morning routine. Instead of a verbal briefing to determine what the floor should work on next, the board answered that question without a meeting.

Family grouping in practice at SQP means that when processing customer orders, the scheduler considers due dates, machine requirements, and material characteristics — diameter, grade, threading — together, and groups jobs that can share a setup. This reduces changeover frequency without sacrificing delivery performance, because the grouping happens within the constraint of the due dates.

The other change Rastko highlighted was floor feedback. When operators complete a job, they update the system. When a job is taking longer than estimated, they flag it. This continuous feedback loop means the estimated times used for future scheduling get more accurate over time — which is how a job shop moves from “I’m not sure when that will be ready” to a specific date it can commit to.

The information underneath the schedule

A scheduling system is only as good as the information it runs on. Rastko was direct about this: the system told them things they hadn’t known before — which machines were waiting for material, which jobs were queued for subcontract work, which setups could run unattended overnight. That information existed on the floor before the system was implemented. It just wasn’t visible to anyone making scheduling decisions.

As Helmuth von Moltke noted, no plan survives contact with the enemy. The value of a good scheduling system isn’t that it makes the perfect plan — it’s that it makes the next adjustment faster and less disruptive than the last one.

Chapters

00:00 — Welcome and Introductions

00:30 — Webinar Objectives and Series Overview — How this session builds on Part 1

02:30 — Traditional vs. High-Mix Manufacturing — Why assembly-line thinking doesn’t apply to job shops

04:30 — The Problem with Static Plans — Why the perfect schedule breaks down the moment work starts

06:37 — Balancing Utilization and Flexibility — The 2x2 framework: from brittle-and-efficient to agile-and-productive

08:30 — Strategies for Balanced Operations — Family-based scheduling, cross-training, and planned flexibility windows

10:31 — Treating Staff as Shared Resources — Moving beyond the one-person, one-machine model

11:30 — Learning from “The Goal” — Theory of Constraints — Why your bottleneck determines total output

14:58 — Managing Dynamic Bottlenecks — How bottlenecks shift with the job mix, and how to track them (Sim Thiam Soon)

17:05 — The Need for Real-Time Information — Why a good schedule is only as good as the data behind it

18:30 — Benefits of Dynamic Scheduling — What changes when the floor can see the plan and the plan can see the floor

19:30 — Key Elements of Change Management — People, process, and technology: what needs to change first

21:29 — Customer Case Study Introduction: SQP Engineering — Context on the business and the problem they were solving

22:10 — SQP’s Journey: From Paper to Digital — How Rastko Petrovic led the transition to infoboard

23:30 — Material Management and Workflow — The colour-coding system and what each status means on the board

25:24 — Maintaining System Currency — Why real-time updates are the hardest discipline to build

28:41 — Family Grouping in Practice — How SQP groups jobs by material and tooling to reduce changeover time

30:56 — Patterns Across Multiple Clients — What Sim Thiam Soon sees consistently across infoWAMS implementations

33:17 — The Quality of Information — Why inaccurate data in a good system is worse than no system at all

35:30 — The Continuous Improvement Journey — How SQP’s time estimates get more accurate over time

36:30 — Series Wrap-up and Next Steps — What Part 3 covers: information flow from sales to dispatch

37:00 — Closing and Appreciation