The research project, carried out in collaboration with the Polytechnic of Milan, has as its object the development of an automatic forecasting solution for future sales, both for consolidated products and for new products.
Abstract, an established partner in the world of Information Technology serving the innovation of market-leading companies, was awarded the Best Paper Award, together with the Data Science team of the Department of Computer Science of the Polytechnic University of Milan, for a research project aimed at creating an automatic sales forecasting solution. The prize was awarded on the occasion of the 24th edition of the International Business Information Systems (BIS) Conference, which took place remotely in Hannover.
The paper, chosen from a shortlist of 64 selected projects, will be published in the BIS volume, considered a reference in the sector of IT solutions to support corporate business. Born from a PoC (Proof of Concept) for the provision of an NPI (New Product Introduction) solution, the project applies Machine Learning to make automatic predictions of future sales, both for consolidated products and for new products in the design phase and not yet placed on the market.
Specifically, the algorithm designed by Abstract is designed to draw on the company's information assets, with the aim of identifying new or existing correlations between the products sold in the past and those that are intended to be launched on the market, based on quantitative and qualitative elements. Through a mapping of sales - distinguished based on the peculiarities of the products (colour, shape, dimensions, components, etc.) - it is possible to obtain an accurate estimate of the sales performance of a new product with similar characteristics, to decide on its launch and predict its potential.
The solution created by Abstract is proposed as an interactive tool to help companies predict market reactions, without replacing the creativity of designers. On the contrary, it represents a valid aid for the design and planning teams, because it optimizes future production in light of market forecasts.
The paper demonstrates how statistical Machine Learning is able to support decisions relating to the design of new products on the basis of information already possessed by each company: the algorithm must, in fact, be fed with data and information from the reference context, making the solution adaptable to any business sector.
The work conducted by Abstract and the Polytechnic of Milan has also received the appreciation of the European Institute of Innovation & Technology, an independent body of the European Union responsible for identifying, co-financing and coordinating the activity of specific "knowledge and innovation communities", such as partnerships between universities, research centers and businesses. The Digital section of the EIT deemed the idea worthy of obtaining European funding.
“The Digital section of the EIT considered the idea worthy of obtaining European funding”, he underlined Edgardo Di Nicola Carena, Head of Data Science at Abstract. “Theory and practice seem like two distant worlds: on the one hand there is the rigor of theoretical and scientific research, on the other the concrete need to implement projects that work and of which the client can appreciate tangible results. By working well together, extraordinary advantages can be achieved”.
