The production lines operate at maximum capacity, order books are full, and the next delivery date is at hand. The last batch is just leaving the coating line when the red light flashes: there are uncoated spots clearly visible on several parts.
Happily, this is not a major issue. The affected parts are stripped, cleaned and sent back to the coating line. Half a day later, the lot is ready for shipment.

Quick fix versus sustainable solution

You probably know similar situations, not only from a production context, but also from other parts of an organisation. A deviation is observed and almost instantaneously someone puts forth a solution – and often this solution remedies the deviation.
Unfortunately, most of these solutions are mere quick fixes, which only address the symptoms. They correct the observed deviation but not the underlying problem which caused the deviation in the first place. Thus the deviation is likely to re-occur later.
In contrast, a sustainable solution does not address the symptoms but the root cause. To address the cause generally requires a bit more time and effort compared to a quick fix. However, it also prevents the same mistake from happening again.

Reasons for a deviation

A deviation in this context means that the actual result does not line up with the expected result. This may happen due to two reasons: either the assumptions on which expectations are based are incorrect or there are flaws in the implementation.

Incorrect assumptions

If you are implementing something new, deviations are often caused by incorrect assumptions. Whether it comes to product development, process optimisation, or the implementation of new tools or methods – you always step out of your experience and explore a new terrain.
In a new environment with only partial information, expectations are mostly based on assumptions, either about the environment, functionalities or causal links. It is important to be aware of these assumptions.

Implementation flaws

If the expected results are not based on assumptions but are derived from a sound understanding of how various factors are correlated and interconnected, deviations often result from flaws in the implementation. It is well-known how to achieve the targeted results, but once or twice this best practice approach was abandoned.
This type of deviation is often observed in areas with clear and stable processes in place like large series production or other frequently repeated processes with low variance.

How to address deviations with long-term success?

If you want to get rid of a deviation once and for all, you should follow a three-step approach:

Analyse the deviation

First of all, you need to understand the type of deviation you are dealing with. Ask yourself whether your expectations are based on an extensive process knowledge or on mere assumptions. If the latter is the case, try to understand which assumptions you have made.

Analyse the root cause

Next, you need to understand the root cause of the deviation. You need to understand which assumptions were incorrect or where the way you implemented certain steps had flaws. Tools like 5-why or Ishikawa are quite useful.

Derive and implement measures

Once you understand the root cause you need to address it. If the cause is a flaw in the implementation, you need to ask: How can I prevent the process from slipping again? The answer usually is to be found in the design of the process or the control mechanism. In this article, we have summarized how you can optimize internal processes.
In the case of incorrect assumptions, you first of all need to replace faulty assumption. Afterwards, you need to assess how, based on your updated knowledge, you ought to proceed.

If you adhere to this procedure and keep in mind our recommendations for optimising processes, you can not only correct deviations and errors, but also fix them sustainably and thus achieve real quality and process improvement. Contact us and we will find a tailored solution for your needs.

On hearing “continuous improvement”, many people thing of the Deming circle or PDCA: plan an improvement, implement it (do), check the results and act upon the them. Once the cycle is completed, it starts anew and thus realises continuous improvement. The second thought, however, may probably be something like “But it’s not that easy!” or “The theory is fine – but it never works that way in real life!” – and considering all the failed initiatives to implement continuous improvement, these thoughts are understandable.

Nevertheless, PDCA is an effective tool to implement continuous improvement. As in many cases, the problem is that this instrument is applied in the wrong way: surveying failed process improvement projects, one comes to the conclusion that mostly the use of tools is responsible for their failure – and not the tools themselves.

Process optimisation as cause-and-effect chain

Process optimisation along the lines of PDCA basically follows the logic of a cause-and-effect chain. An existing process is deliberately modified (= cause) and as a result the efficiency or effectiveness of the process improves (= effect). This sounds trivial but, far from it, actually significantly impacts the approach to process optimisation.

Causal relationship between cause and effect

If process optimisation follows the logic of cause-and-effect chains, then there is a causal relationship between cause and effect. This relationship is created by the process which is to be improved and the interfaces with other processes.

Generally, these causal relationships are not just linear dependencies but form complex networks. These networks are normally not limited to the process to be improved but linked to other processes and activities by interfaces.

Explicability of the effect

Another consequence of process improvement as a cause-and-effect chain is that you can explain the effect of any process modification. Causal relationship provides a rational explanation why a certain effect results from a given cause.

The rational explanation may be used in two ways. If the causal network is known, the effect of any modification can be predicted. Likewise, if both cause and effect are known, the causal network can be reconstructed.

Optimising causal networks based on PDCA

Given the implications of process optimization as a cause-and-effect chain, PDCA is ideal for continuous improvement. In this context, the respective steps are:


Before first applying PDCA, the optimization target needs to be defined (see „How to optimise internal business processes“ [Link:]). The target should be clearly defined and quantified whenever possible.


Based on the current understanding of the process, i.e. its causal network, an optimisation step is worked out which modifies to process depending on the defined target. Having done so, both the necessary measure (= cause) and the expected outcome (= effect) have to be described.


The described step is implemented, with a constant monitoring of the resulting effects on the process.


Once the optimisation step is implemented, the actual effects are compared with the expected ones. If the observed results match those expected in the plan-step, the PDCA iteration is completed. In case the overall target is met, the optimisation cycle is completed, otherwise another iteration begins (i.e. plan).

If the observed results were not expected, the current process knowledge needs to be adjusted. To allow for such an adjustment, it is essential to understand why the implemented change resulted in the observed effect.


Once the causal relationship between the implemented measure and the observed effect is established, the knowledge about the process needs to be updated accordingly. The updated knowledge will then serve as a basis for another PDCA iteration.

Benefits from PDCA process optimisation

Using PDCA for process optimisation requires thorough application, which may prove difficult in busy day-to-day operations. Most of all, the analysis of unexpected effects is readily skipped in favour of quickly implemented fixes.

However, by sticking to the steps described above, you will win twice: First, you will achieve the set targets for process improvements. Second, your knowledge about relevant processes will increase.