Kaizen

The operational rhythm follows a repeatable sequence:

1. See: Operators identify an obstacle, variation, safety issue, or non-value-adding task during daily production. To learn how to classify physical and procedural wastes, refer to The seven wastes.
2. Try: The team formulates a hypothesis and implements a small countermeasure on the shift.
3. Check: The team measures performance metrics before and after the trial across identical sample sizes to verify if the countermeasure caused an improvement.
4. Standardize: If the trial is successful, the team updates process sheets, work instruction documents, and visual controls. If the trial fails, the team reverses the change and tests an alternative countermeasure.
5. Repeat: The updated standard becomes the baseline for the next cycle of analysis.

For instructions on executing this cycle in structured operational settings, see the guide to Run a PDCA cycle end to end. When balancing cycle times across multiple process steps, teams can analyze work element allocations using a Yamazumi chart builder.

In a traditional suggestion-box model (Programme A), twelve employees submit 300 suggestions annually, averaging 25 proposals per person. In this model:

• Ideas are evaluated by a monthly committee with a median review time of 31 days.
• Standard work updates are left to a centralized continuous improvement department.
• Under this structure, only 8 percent of ideas reach implementation (24 changes).
• Only 38 percent of implemented changes survive six months, yielding 9.1 sustained kaizens per year, or 0.76 per person.

In an active, shop-floor board system (Programme B), twelve employees submit 90 suggestions annually, averaging 7.5 proposals per person. In this model:

• Ideas are reviewed during daily team huddles with a median response time of 1 day.
• The frontline team updates its own standard work during the same shift.
• As a result, 78 percent of ideas are implemented (70 changes).
• 62 percent of implemented changes remain in place at six months, yielding 43.5 sustained kaizens per year, or 3.6 per person.

Programme A converts raw ideas into lasting improvements at a rate calculated by multiplying the implementation rate by the survival rate: multiply 0.08 by 0.38 to get 0.0304 (3.04 percent). Programme B converts ideas at a rate calculated by multiplying 0.78 by 0.62 to get 0.4836 (48.36 percent). This represents a 16-fold difference in conversion efficiency.

For Programme A to match Programme B's annual output of 43.5 sustained kaizens without altering its centralized process, its workers would need to submit 1,431 ideas annually. This equals 119 suggestion cards per person each year, or roughly one proposal every other working day. When feedback takes 31 days, worker momentum disappears, whereas a single-day response window enables workers to test countermeasures immediately while operational context remains fresh.

Consider the cycle times observed when attempting to eliminate 9.0 seconds of walking time at a workstation:

Cycle 1: Repositioning existing bins

• Plan: Identify excess walking motion timed at 9.0 seconds per unit across 20 cycles. The team proposes moving existing parts bins closer to the front edge of the assembly bench.
• Do: Shift the parts bins forward during the scheduled changeover break.
• Check: Re-time the workstation across 20 cycles. The average time measures 44.0 seconds, representing a minor reduction of 2.0 seconds instead of the target 9.0 seconds.
• Act: Recognize that the layout modification fails to eliminate reach-and-twist motions. The team decides not to standardize this configuration because moving the bins crowded the assembly surface.

Cycle 2: Gravity-fed chutes

• Plan: Design a gravity chute from the supply rack over the bench, keeping the base bins fixed while feeding parts directly to the point of use.
• Do: Install a low-cost cardboard and pipe prototype chute for the afternoon production run.
• Check: Re-time across 20 cycles. Average station time drops to 37.0 seconds, successfully securing the target 9.0-second reduction.
• Act: Standardize the gravity chute layout, replace the prototype with permanent materials, update standard work sheets, and train operators across all shifts.

The failure of Cycle 1 is an essential part of the scientific method. A 2.0-second variance on a 46.0-second operation falls within normal statistical process variation; recording it as an improvement would introduce ergonomic strain without meaningful cycle gains.

To ensure the Check phase remains reliable, two practices are required:

1. Sample consistency: Measure identical sample sizes before and after changes (for example, comparing 20 cycles against 20 cycles rather than a 5-cycle baseline against a 20-cycle trial).
2. Structured documentation: For complex or cross-functional issues, using structured A3 problem solving formats helps document before-and-after baseline metrics alongside root-cause analysis on a single sheet.

What distinguishes a true kaizen from a standard workplace suggestion?

A workplace suggestion is simply an unverified idea, whereas a true kaizen requires frontline ownership, a verified measurable improvement, and sustained documentation in standard work instructions. Tracking raw idea submissions creates a false impression of progress if those ideas are never implemented or sustained. A proposal only becomes a kaizen once operators test it, measure a positive change against baseline performance, and confirm its continued use through operational audits.

Why do frontline management boards produce more sustained kaizens than centralized suggestion boxes?

Centralized suggestion systems rely on committee reviews with long turnaround times, such as a 31-day latency, which causes worker momentum to fade and yields implementation rates around 8 percent. In contrast, active shop-floor boards review proposals during daily huddles with a single-day response window, allowing operators to test ideas immediately while context is fresh. This rapid frontline feedback loop converts submitted ideas into sustained operational gains at a rate up to 16 times higher than centralized programmes.

Why does a verified kaizen countermeasure decay without immediate standard work updates?

When an improved method is not immediately formalized in standard work instructions, natural process variation and shift turnover lead workers to instinctively revert to legacy habits. For example, an undocumented assembly change can lose most of its cycle time reduction within 90 days as operators rotate through the station. Updating standard work on the same shift provides an authoritative baseline that preserves labor capacity gains across all operating teams.

What causes a kaizen event to fail after initial completion?

A five-day kaizen event fails when an organization treats the workshop as a one-time project and neglects post-event follow-up. Without structured governance, up to 40 percent of secondary action items decay after the initial event concludes. Sustaining improvements requires leaders to maintain a weekly review cadence and audit open action items at 30, 60, and 90 days.

Why are unsuccessful trials an expected part of the kaizen process?

Kaizen relies on the scientific Plan-Do-Check-Act cycle to differentiate genuine process improvements from baseline operational noise. Testing small, low-cost countermeasures allows teams to observe actual effects at the process location and quickly discard changes that introduce ergonomic strain or fail to meet performance targets. Rejecting ineffective trials prevents unhelpful modifications from being standardized and directs focus toward countermeasures that deliver verified gains.