DIGGING UP YOUR OWN DATA TREASURE
CAD data is a treasure trove of information. Shouldcosting analyses this information to identify potential savings in purchasing, or to calculate costs for new components. QUO is one of the companies that relies specifically on these analyses for certain projects. Hans-Peter Gysel, CEO of Shouldcosting, explains how businesses can leverage big data and how product development can benefit from it.
Mr Gysel, you use Big Data to reduce manufacturing costs. How does it work?
3D CAD files and 2D drawings contain an enormous amount of cost-relevant information – the shape of a component, its size, the number of holes, material properties, surface coatings, etc. We feed this data fully automatically into our system, search it for cost-relevant information, and structure it. This data mining strategy enables us to quickly determine how many similar or identical parts a company uses or buys.
How do you compare the costs?
In order to do this, our algorithms calculate correlations between the technical specifications of the components, the batch sizes, the prices, or the countries where the components are produced. If there are any outliers in the data, we take a closer look. For example, we may find large price differences between similar products from the same supplier, or between similar products from different countries. This is where companies can make cost savings.
Companies entrust sensitive information to you. How do you handle this sensitive information?
Data protection is a top priority for us. All our systems are triple-encrypted and we conclude a non-disclosure agreement with each of our customers. In some cases, we also analyse the data directly on the customer’s premises, which means the data never leaves the company. With the customer’s permission, we can also use the data for our reference database. However, when we do this, no drawings or pictures are stored – instead we only store numbers in the form of digital, anonymous fingerprints of the items that have been analysed.
QUO has also made use of your expertise. One example is the project for Bibliotheca, in which the manufacturing costs were reduced by 20% and the number of parts required by 30%. How was this achieved?
In this case, QUO reduced costs significantly by combining different devices with a wide variety of components into a single device concept with a modular architecture. An algorithm can’t come up with such a solution – only resourceful minds can. We contribute to this kind of project by determining which components have the greatest savings potential. Furthermore, after we perform data mining for a company, we have access to pricing information from all of their suppliers. In QUO’s case, we were able to use this information to quickly calculate target costs for new components (predictive costing).
You also work with QUO as a partner in other projects. How do you and QUO complement each other?
We are ideal partners and we complement each other very well. QUO looks at a machine or product group in depth and optimises the entire design to reduce the number of assemblies, simplify the assembly process, or improve user-friendliness.
Our algorithms perform an analysis across the entire company, which allows us to quickly compare different manufacturing variants, or assess which supplier or which country can produce a component for which price. This makes it easier to control subsequent manufacturing costs during product development. Our analysis of large amounts of data enables us to uncover savings potential within a very short time – including savings potential that nobody would ever have imagined.