13.09.17, 08 Uhr
So1 - Uncovering the DNA of Shopping
So1 is one of the driving forces of retailer digitalization. Deeply understanding what consumers want, and then adapting offers accordingly, make individual promotions so powerful. The first step is to uncover the DNA of shopping. Whether or not a consumer will buy a product is a function of her or his (mostly unobserved):
- preferences for product attributes in general,
- the specific attributes of the product,
- his purchase probability at a given time and
- the resulting willingness-to-pay taking into consideration his alternative actions, plus
- the on-shelf availability of the product.
From millions of baskets, the So1 Engine understands how consumers perceive products and their properties. Based on deep learning algorithms, it creates a 800+ dimensional space of latent product properties. These properties can capture market structures, based on
- category, e.g. dog food
- meta-category, e.g. home care
- consumer target audience, e.g. products for person suffering from gluten intolerance
- intended use, e.g. breakfast
- cooking/recipes, e.g. Mexican dishes
In a second step, the individual consumer is placed inside the 800+ dimensional space. Based on the distance between each individual consumer and the topology of products So1 derives the willingness-to-pay. While selecting the relevant product per consumer delivers 60% of So1's efficiency, selecting the right discount and purchase probability delivers the remaining 40%.
So1's AI Targeting Engine is able to answer the following exemplary questions:
- Which products do consumers like (without having observed them buying those)?
- What specific properties do consumers like in products?
- What are consumers willing to pay for a product?
- How likely is a purchase of a product at a given time?
- Which consumers will be attracted by a product launch?
- Which products are seen as substitutes in the eyes of a consumer?
- Which consumers are likely to increase their frequency with a brand or retailer?
- The combination of which offers will deliver the greatest value to a retailer?
- What is the best time to distribute a specific offer?
address: Oderberger Str. 44, 10435 Berlin, Germany
phone: +49 (0)30 208 987 30
For more information, please visit our homepage and blog on Artificial Intelligence & Retail:
Die Online-Pressefächer sind ein gemeinsamer Service der Koelnmesse und dem Originaltext-Pressedienst press1.de Verantwortlich für den Inhalt der Pressemitteilungen sind die jeweils herausgebenden Unternehmen. Diese stellen Koelnmesse von jeglichen Ansprüchen Dritter, die im Zusammenhang mit dem Gebrauch der Online-Pressefächer entstehen, frei. Diese Freistellung gilt insbesondere auch für Verstöße gegen das Urheber- und Markenrecht. Sie besteht auch dann fort, wenn die betroffene Pressemitteilung bereits aus den Online-Pressefächern herausgenommen worden ist.