Close this search box.
Close this search box.

Case History

Six Sigma Strategies and DFx Method in Product Development

Early identification of causes of faults and failures in the product to improve quality and contain costs.
six sigma, 6 sigma, DFM, Design For Manufacturing, capacità di processo


Family business, 95M€ turnover, about 300 employees


The company

The company is a family business with 95M€ turnover and about 300 employees, present in the market since 1960. Strongly innovative, with over 200 patents filed, it specializes in the development, industrialization and production of molded plastic components used in containers in the cosmetics market.


The challenge

The production was slowed down by frequent requests to the design and quality department for production approvals and waivers due to non-standard tolerance of some components that could actually be used without affecting the finished product. The process was holding up production, increasing its time-length and, consequently, also its cost.

In order to limit this continuous bounce-back between design and production, the method of defining quality requirements needed to be overhauled.


The solution

The proposed solution was to introduce Six Sigma strategies, which are functional in improving the efficiency of the link between design and production: in fact, they allow early identification and removal of causes, defects and failures in the product. The Six Sigma techniques model the system at the design stage, using either historical data or estimated data derived from the company know-how, whose importance is always too underestimated.

n particular, the Design For Manufacturing (DFM) method was chosen.

Embedded early in the product development process, the DFM allows the optimal model to be built, with the indication, for each component, of tolerances large enough to be feasible and at the same time adequately contained to allow excellent functionality. The result is a better quality product at a lower cost.


The operational phase


Changing the determination of tolerances, passing from a deterministic to a statistical approach, was useful to design thinking about the world as it really is, and not as we would like it to be!”


The DFM was focused on the statistical analysis of mechanical tolerances for each part of the product (stack-up analysis).

The first step was to have a conversation with the end customer on the one hand, analyzing what the actual dimensional characteristics were that were needed to ensure that the product works; and with the production quality department on the other, gathering precise feedback on how the tolerance deviations were impacting the effective quality of the product and machining. We found that, in many cases, even with seemingly out-of-tolerance components, the product was performing excellently.

The second step brought together the design with the quality body to re-discuss and validate the actual CTQs (Critical To Quality), i.e., the critical parameters for quality that are really needed to ensure the functionality of the product. In this way, it we realized that not all dimensions are important and essential for functionality: but there are some that are absolutely mandatory and sometimes were missed. This work was validated through the construction of a functional model that made it possible to identify the key dimensions on the dimensional tolerance of which to collect actual data: through the statistical analysis of these data, it was possible to identify which tolerance ranges were effectively critical and which were not, and it was realized that, contrary to the historical beliefs of the designers, it is not necessary to impose particularly tight tolerances in order to make products work properly.

The third step was to develop a model for the design department, standardized and replicable, that would allow the maximum tolerance width for each component to be uniquely defined and characterized with the appropriate capability, in relation to the functionality on the one hand and the characteristics of the production process on the other.

The main difficulty one encounters with this methodology is the resistance to setting aside confidence with minimal tolerances by the design. The second critical element is data collection, on which the 6 Sigma model is based: in fact, there is a shift from a deterministic approach, typical of designing, to a statistical approach, more anchored to reality, but in need of an information base on which to be developed. The good news is that, in practice, any company can, with relatively little effort, collect useful data to make these models work!


The results

The codesign process led to:

  • 45% fewer irrelevant measures applied on projects
  • 50% average increase in tolerance range

Tolerances, brought to more realistic and functional measurements, allowed for the acceleration of the fine-tuning of the production processes, without inducing unnecessary revisions of stages or equipment that added nothing to the quality of the product.


It is the precise definition of the expected tolerance performance for the entire system that guarantees the performance desired by the customer.”


Roberto Malaguti – Partner ŌdeXa


Are you interested in learning more about the topic?


form business case

You might also be interested in