SMED: How to Reduce Changeover Time and Improve OEE with Data

SMED (Single-Minute Exchange of Die) is the most powerful lean manufacturing methodology for reducing setup and changeover time — and one of the highest-ROI OEE improvement levers available to manufacturing operations. Developed by Shigeo Shingo at Toyota, SMED aims to reduce equipment changeover time to less than 10 minutes (“single minute” means single-digit minutes, not one minute). In practice, structured SMED implementation typically achieves 30 to 60% changeover time reduction — a gain that translates directly into OEE availability improvement and production flexibility. This guide explains how SMED works, where data accelerates results and how TeepTrak supports systematic changeover time reduction.

Why Changeover Time Matters for OEE

In most manufacturing environments, changeover and setup time is either the largest or second-largest OEE availability loss. Every minute the equipment is stopped for a product change is a minute of lost production capacity. On a production line running 3 changeovers per shift with an average changeover time of 45 minutes, the total changeover loss is 135 minutes — nearly 30% of an 8-hour shift. If SMED reduces that to 25 minutes per changeover, the gain is 60 minutes per shift of additional production time, equivalent to a 12.5% OEE improvement from this loss category alone.

Beyond the direct OEE impact, reducing changeover time has a strategic dimension: it enables smaller production batches, more frequent product switching and faster response to demand changes. Factories with short changeover times can run what customers need, when they need it, without accumulating large inventory buffers to compensate for long, infrequent production runs.

The SMED Methodology: Three Steps

Step 1: Separate internal and external setup activities. Internal setup activities are those that can only be performed when the machine is stopped (removing old tooling, installing new tooling, making adjustments). External setup activities are those that can be performed while the machine is still running the previous production run (preparing new tooling at the side of the machine, pre-heating dies, pre-staging materials). The most immediate SMED gain comes from converting internal activities to external — typically achievable without any equipment modification. This alone can reduce changeover time by 30 to 40%.

Step 2: Streamline remaining internal activities. Once internal activities are minimised, the remaining internal changeover time is reduced by standardising procedures, improving tooling design, eliminating adjustments and reducing fastening operations. This step often requires some investment in tooling or fixtures but delivers the largest changeover time reductions.

Step 3: Eliminate adjustments and trial runs. Many changeovers include a series of trial runs and micro-adjustments after the new tooling is installed — the machine is started, a sample part is produced, adjustments are made, another sample, more adjustments, until the process is stable. Eliminating these trial runs through standardised tooling reference positions and first-piece guarantee procedures is the final SMED lever.

How Data Accelerates SMED Results

Traditional SMED implementation relies on direct observation: a team stands at the machine with a stopwatch and records every activity during a changeover. This is time-consuming, subject to observer bias and only captures a single changeover event. The resulting time study may not represent typical changeover performance — especially if operators perform differently when being observed.

TeepTrak transforms SMED implementation by providing automated changeover time data across every changeover event, not just observed ones. Key data outputs that accelerate SMED:

Actual vs standard changeover time per product transition: TeepTrak measures the exact duration from last good part of the previous run to first good part of the new run, for every changeover, automatically. Plotting these actual durations against the standard time per product transition immediately reveals which transitions consistently overrun, by how much and with what variability. The highest-priority SMED targets are the transitions with the highest actual-vs-standard gap.

Changeover time by operator: The same product transition performed by different operators often shows 20 to 50% variation in changeover duration. JEMBA AI identifies operator-specific patterns — whether certain operators consistently outperform the standard (indicating best practices to standardise) or underperform (indicating training needs).

Startup loss data: TeepTrak tracks not just the changeover time but also the startup loss — the time and output quality from first part to first confirmed good part. A changeover that takes 20 minutes but requires 15 minutes of startup adjustment has a 35-minute total production impact. SMED improvement that reduces changeover time without addressing startup loss captures only part of the opportunity.

Changeover frequency analysis: Which product transitions happen most frequently? Which have the highest total time impact (frequency × average duration)? TeepTrak generates this Pareto automatically, enabling SMED teams to focus on the transitions that will deliver the highest overall OEE impact rather than the longest individual changeovers.

TeepTrak SMED Module in Practice

TeepTrak PaceTrak module provides dedicated SMED tracking alongside standard OEE monitoring. Changeover events are automatically detected (no operator marking required), timed and categorised. Production managers can access real-time changeover duration on any line, historical Pareto of changeover time by product transition and machine, and trend analysis showing whether SMED improvements are being sustained over time.

JEMBA AI monitors changeover performance continuously — alerting when a specific transition’s average time begins increasing (indicating process drift or new operator without training), and identifying patterns in changeover variability that suggest specific improvement opportunities.

FAQ

What does SMED stand for and what does it mean?

SMED stands for Single-Minute Exchange of Die — “die” referring to manufacturing dies and tooling in the press and stamping context where the methodology originated. “Single minute” means single-digit minutes, not literally one minute. The goal is to reduce changeover time to less than 10 minutes. In practice, structured SMED typically achieves 30 to 60% reduction in changeover time from baseline, regardless of the starting duration.

How long does a SMED project take?

A focused SMED project on a single product transition typically delivers initial results within 2 to 4 weeks — one week of baseline measurement and analysis, one week of improvement design and implementation, and one to two weeks of verification and standardisation. With TeepTrak providing automated baseline data (eliminating the manual time study phase), the total project duration is typically shorter and the improvement actions are better targeted.

What is the difference between SMED and changeover time?

Changeover time is the duration of a product or tooling transition on a production line — from last good part of the previous run to first good part of the new run. SMED is the methodology for reducing that changeover time. SMED provides the structured analytical and improvement framework; changeover time is the metric that SMED aims to reduce. TeepTrak measures changeover time automatically and provides the data that SMED improvement teams need to identify and prioritise improvement opportunities.

Can SMED be applied to cleaning and CIP cycles in food manufacturing?

Yes. In food and beverage manufacturing, CIP (Clean In Place) cycles and allergen changeovers are the dominant setup and adjustment losses. SMED principles apply directly: separating preparatory activities (cleaning solution preparation, documentation) from internal CIP time, standardising CIP procedures per transition type, and eliminating adjustments in post-CIP startup procedures. TeepTrak tracks CIP actual vs standard duration automatically and JEMBA AI identifies which product transitions consistently generate CIP overruns — the starting point for food-specific SMED analysis.

Measure and reduce your changeover time — free 48h POC

See how TeepTrak SMED tracking has reduced changeover time across manufacturing sectors. Visit our customer success stories.