Issue link: https://iconnect007.uberflip.com/i/1426508
NOVEMBER 2021 I DESIGN007 MAGAZINE 59 according to which we can base a model that can predict the actual cost of a product. If, say, we know from experience, the failure rate of a certain category of products is 20%, that means the quote should rise accordingly. Failure ratio is an integral part of any future quote. When you plan a production route based, for example, on a set of 100 actions, and in the actual process we realize there's a failure which calls for repair and adds yet one or more actions to the process, that means the quote should be higher. at requires digitally signed and docu- mented actions. Each action has an owner, a start, and a completion time report, and of course the indication of the machine on which it was produced. is is the only way that we can enable the system to add data, analyze it, and integrate it with other pieces of informa- tion already deployed. Data is processed daily. Some models undergo the same procedure on an hourly basis. e result is the ability to predict not only the failure ratio but also lead time, which also has bearings on the price. e system is based on the business intelligence soware QlikView, which processes the infor- mation. e data is saved for a long period and can be used whenever necessary. We have faced the same challenge with our customers over the years. Our new pricing sys- tem, JUMP, was designed especially for this purpose. e idea was to enable customers to get an accurate quote as painlessly and as fast as possible. is big-database analysis at the heart of the system runs scripts that can predict, in the early RFQ stage, the final product price and, indicate their lead-time and level of com- plexity. Our pricing fit model is based on more than 6,000 different part numbers that have completed the full engineering process, and includes their BOM, routings, and material clas- sifications. e result is a unique and extremely reliable system with a very high R square of 0.98. e unique manner in which data is man- aged in the system allows for statistical identi- fication of various raw material combinations, Figure 2: Big data analyses helped us to build a script that predicts, in early RFQ stage, the final product price based on a very short classifications questionnaire.