Back to Blog
Smart Factory

The Time Lost in Changeovers: Digital Changeover Optimization for High-Mix, Low-Volume Manufacturing

As lot sizes shrink, capacity is determined by changeover time rather than machine speed. This post shows how to make invisible changeover losses measurable — hidden inside OEE planned downtime — and cut them 20-30% by fixing production sequencing and preparation work together.

POLYGLOTSOFT Tech Team2026-08-278 min read0
High-Mix Low-VolumeChangeoverSMEDProduction PlanningMES

A Market Where Orders Keep Getting Smaller

Work orders on the 2026 factory floor are far thinner than they used to be. Tariff volatility and demand uncertainty have broken large purchase orders into fragments, and a line that once ran 100,000 units in a single batch now absorbs twenty separate orders of 5,000.

When lot sizes shrink, the variable that determines capacity changes. With a 50,000-unit lot, a two-hour changeover accounts for barely 3% of the time. Split that into 2,000-unit lots and the same two hours consume more than 40% of total operating time. At that point capacity is decided not by machine speed but by changeover time — which is exactly why more manufacturers are betting on agility rather than large-scale automation spending.

Why Changeover Losses Stay Invisible

Most changeover loss disappears into the planned-downtime bucket of OEE. The moment it is labeled "planned setup time," it drops off the improvement list.

Break changeover into these five segments and the loss becomes visible:

  • Die and tooling preparation: half the time is walking to storage and searching
  • Material swap and return: waiting on leftover-quantity reconciliation
  • Cleaning and purging: spikes whenever color or material changes
  • First-article inspection: waiting for an inspector takes longer than measuring
  • Approval wait: the dead time until a quality manager signs off
  • Measure it in the field and you routinely find the same item pair taking 40 minutes for one operator and 95 minutes for another. Most of that gap comes not from manual speed but from when the materials arrived and how long approval took.

    Measurement: Turning Changeover Time Into Data

  • Automatic segmentation: let the system carve out the interval from the previous work order's completion to the first good-part report on the next order. The instant you ask operators to key it in manually, the data loses credibility.
  • Internal vs. external classification: record separately the work that requires the machine to be stopped (internal) and the work that can be finished while it still runs (external). It is the most intuitive metric for showing how much room for improvement exists.
  • Item-pair tables: accumulate average, median, and maximum values for every from-to combination. Three months is usually enough for the distribution of your top pairs to stabilize.
  • Improvement: Change Planning and the Floor Together

  • Sequence by cheapest changeover: feed the item-pair table into production sequencing and orders sharing a color family or a die group together naturally. Across a typical week of 20 orders, resequencing alone has cut total changeover time by 20-30%.
  • Issue preparation as its own task: release material and tooling preparation for the next order as a separate work task handled while the current order is still running. It is the cheapest way to convert internal work into external work.
  • Shorten inspection and approval: digitize first-article inspection criteria as an electronic check sheet and add mobile approval. Approval waits of 20-30 minutes shrink to a few minutes.
  • Distribute the standard procedure: turn the fastest operator's method into a tablet checklist, deploy it, and verify the effect with actual performance data.
  • Common Misconceptions in High-Mix Production

    "Automation will fix this." Dedicated equipment raises throughput for a single item but actually reduces changeover flexibility. On a line whose product mix keeps shifting, general-purpose equipment plus a fast changeover system is often the better bet.

    Planning and execution living apart. If the sequence ERP produces carries no changeover-time information, the floor simply rebuilds it in whatever order suits them. The plan was not ignored — it was a plan worth ignoring.

    A Phased Rollout Roadmap

  • Start measuring one line and its top 10 item pairs (4-6 weeks)
  • Apply sequence optimization and compare before and after
  • Split out preparation work and issue it in advance
  • Roll the validated standard procedure out across all lines
  • How POLYGLOTSOFT Approaches It

    We do not rip out your existing MES or ERP. We leave the running systems in place and layer a changeover loss measurement module on top, starting with nothing more than building the item-pair changeover table. Once the data accumulates, we attach the sequencing logic and preparation-task issuing on top of it.

    Because the product mix keeps changing, sequencing logic is never something you build once and walk away from. POLYGLOTSOFT's subscription development model means the logic gets revised and verified every time line conditions shift, so improvement does not stop after the initial build. If changeover time is quietly draining your capacity, tell us about your current line configuration through our [contact page](/support/contact) and we will map out the measurement scope and expected impact with you.

    Need Technical Consultation?

    Our expert consultants in smart factory, AI, and logistics automation will analyze your requirements.

    Request Free Consultation