Digital Andon connects an abnormality signal to a defined response process. Its Lean foundation remains jidoka: Toyota describes operators stopping a line when an abnormality occurs, with an Andon light calling the person responsible, on global.toyota. A digital implementation needs to specify what happens between that signal and verified recovery.
Para los profesionales que estén considerando los programas de certificación de Lean Six Sigma International, Programas de certificación, digital Andon offers a bounded project: measure delays against an agreed response standard. The team can then test whether changes prevent recurrence and use the findings to write the software specification.
Jidoka sets the requirements for digital Andon
Toyota’s explanation of jidoka connects abnormality detection with stopping production to prevent defective output from continuing downstream, as described on global.toyota. In a digital Andon design, the plan de control must specify which conditions require a stop and who can authorise restart. A notification deadline must never become permission to continue producing suspect material.
A gemba walk lets the team trace an abnormality from detection to recovery. The record needs to show who received the signal and how they accepted responsibility. Evidence supporting restart deserves separate scrutiny. These observations establish the requirements for trabajo estandarizado before the team selects screens or messaging channels.
A published escalation system shows the routing decision
István Gödri’s 2024 paper in Periodica Polytechnica Mechanical Engineering, available at pp.bme.hu, describes a digital escalation system in a high-mix, low-volume manufacturing plant. Operators issued timestamped electronic tickets to specialist responders. Defined time limits governed when unresolved issues moved to higher organisational levels. The design extends Lean gestión visual by assigning responsibility for responding to each signal.
The paper distinguishes material shortages caused by delayed internal delivery from material awaiting incoming inspection, as documented at pp.bme.hu. That distinction has a direct consequence for ticket routing: “Material missing” does not identify the function able to investigate. The ticket needs clarification before assignment, while preserving the operator’s original observation.
A separate study published in Computers in Industry in October 2021 examined escalation management at WITTENSTEIN. Researchers found that responders could be unavailable or lack the skills needed for the disruption, according to the study on sciencedirect.com. The project’s DMAIC analysis therefore needs to separate notification delay from incorrect assignment. Sending an alert sooner does little to resolve a skills mismatch.
Design the response sequence before configuring alerts
The proposed digital Andon pilot sequence below separates acceptance of responsibility from verified recovery. It also gives recurring events a route into análisis de causa raíz.
Detect abnormality
Apply stop or containment rule
Route the event
Identify the responsible role
Confirm ownership
Escalate if acceptance is overdue
Verify recovery
Apply defined restart criteria
Review recurrence
Assign an A3 where justified
For each event category, the Process Owner needs an agreed acknowledgement deadline and a separate escalation condition for unresolved problems. An unavailable responder and a shift handover are useful pilot tests. In both cases, every open event must retain an identifiable owner.
The initial Andon record needs to be short enough to complete at the workstation. Suggested fields are location, observed abnormality, detection time and containment status. Confirmed cause belongs in the record after investigation, with evidence supporting the conclusion.
Measure acknowledgement separately from recovery
The DMAIC measurement plan needs timestamp definitions before the team compares performance. The pilot can use the following intervals:
- Acceptance delay: event creation to acknowledgement by the responsible responder.
- Intervention delay: acknowledgement to the start of corrective work.
- Recovery duration: event creation to verified restoration of the process.
Comparisons by event category and shift depend on a baseline checked for missing timestamps. Recurrence after closure provides a medida de equilibrio, using a review window agreed for the process. If ticket closure gets faster while the same abnormality returns, the team needs to examine the closure criteria.
Gödri’s study analysed reaction times and repeated deviations, and describes changes in the categories covered by the system over time, as reported at pp.bme.hu. A Lean Six Sigma team needs a record of those scope changes alongside its ticket counts. Adding incoming-inspection delays to a system that previously captured only internal-delivery delays, for example, changes the basis of comparison.
Use repeated Andon events to select A3 work
At the pilot review, a gráfico de Pareto of recovery time by confirmed problem category shows where recovery consumes the most time. Recurrence records help identify problems that keep requiring containment. An A3 problem-solving assignment gives that investigation an owner, with a defined test to establish whether the countermeasure prevents recurrence.
The WITTENSTEIN study also explored predictive responder selection and warned that incorrect predictions can cause wrong dispatches, as described on sciencedirect.com. Before adding that capability, a Lean Six Sigma pilot needs a measurable assignment baseline and a manual reassignment route. Misrouted events belong in response-time reporting because their delays are part of the process being measured.
El programa de Lean Six Sigma International Programas de certificación cover DMAIC under its LSSx.0 standard. Green Belt coverage includes measurement system analysis and statistical process control, as detailed on leansixsigma.org. The level selected should match the analytical responsibility assigned in the Andon improvement project.
