KI-gestütztes Forecasting im B2C-E-Commerce

: Potenziale, Herausforderungen und Handlungsempfehlungen am Beispiel der Beauty- und Personal-Care-Branche

Abstract

This master's thesis deals with the use of AI-supported forecasting in B2C-E-Commerce in the beauty and personal care industry. The basis is an online retailer survey conducted in 2024 with 624 participants from Switzerland, Austria, and Germany. This shows that AI technologies are increasingly used in e-commerce, but mainly in areas such as marketing or content creation. In forecasting, however, only 4 % of the companies surveyed already use AI-supported methods, while only 11 % plan to use them in the future.2 This comparatively low level of use in relation to other areas of application forms the starting point for this thesis. The aim is therefore to identify the current situation in the DACH region and determine the factors that are decisive for the use of AI-supported forecasting. To answer this question, a literature review is conducted at the outset, in which traditional forecasting methods and AI-based approaches are analyzed and compared. Subsequently, guided interviews are conducted with experts from the beauty and personal care industry to gain practical insights. The results show that traditional methods still outweigh AI-based approaches, which are currently either only being tested or are not yet considered relevant. However, experts consider them to be promising for the future. The key potential lies in the increased accuracy of forecasts, the integration of external influencing factors, and the possibility of real-time forecasts. At the same time, there are challenges, especially with regard to data quality, high modeling costs, limited interpretability, and organizational factors such as a shortage of skilled workers or acceptance problems. Based on these theoretical and practical findings, recommendations for action are formulated in four areas. These include strategic, organizational, technical and industry-specific measures. The findings of this thesis illustrate that AI-supported forecasting in the beauty and personal care industry is still in its early stages, but is expected to become increasingly important in the long term. In the short term, companies are likely to continue to rely on traditional methods and expand them through the use of multivariate methods in order to include additional influencing factors. This will allow them to improve their database and create the organizational conditions necessary before AI-supported forecasting is progressively considered.

Date of Award	2025
Original language	German (Austria)
Awarding Institution	Johannes Kepler University Linz
Supervisor	Patrick Brandtner (Supervisor)