Quality Improvement

Many small and midsize organizations in don’t have a robust quality systems! It is imperative to put one in place. A good system includes, among others:

Reviewing designs and validating that they are fit for production;
Ensuring that standard work instructions are relevant and are followed by workers;
Identifying and segregating non-conforming components and products;
Implementing preventive and corrective actions

Other factories have a system in place, but don’t respect it fully. The problem might be:

The inadequacy of that system – it needs some adaptation or simplification.
Lack of a layered process audit
A lack of discipline from the staff etc

Coaching managers to maintain staff discipline

Managers need to get out of their office and walk the production floor many times during a shift. This coupled with visual management tools; performance boards filled out every hour, for example allows leaders to have an idea of what is going on. It is an opportunity to reinforce the importance of discipline and to put pressure on supervisors and team leaders when necessary. Among others, the following rules must be followed by everybody:

We do not accept bad work
We do not do bad work
We do not pass on bad work

Formalizing “standard work” and training the staff

For each task, a simple and visual guide, a one point lesson should be available to the staff. It should show the current best way to perform that task and should emphasize critical points (for quality and safety).

Mistake proofing processes against human errors

Humans make mistakes … But the most frequent human mistakes can be identified and addressed with three mistake proofing processes:

Prevented right where they appeared. Example: a part that can only be inserted one way into a fixture.
Detected right when they appear. Example: a sensor detects that only 2 out of 3 holes were drilled, and shows a red blinking light.
Detected after they appear. Example: the next operator inspecting each part.

Analyzing processes for other types of problems

When mistake proofing is not enough or not possible, we favor two other types of approaches.

Root cause analysis, followed by corrective actions and follow up. It is important to trigger this action as early as possible. See 4 quality improvement tools that we use to that effect.
DOE (design of experiments), when many variables may impact the final result and the cause of the problems is not clear.