Data strategy for the introduction of artificial intelligence (AI)
Our client, a sucessful retailing platform wanted to offer their B2B customers additional assistance systems after the third round of financing. niologic moderated the strategy workshop, supported the roadmapping the data strategy and the acquisition of new team members.
Challenge
The client wanted to channel the investments from another round of financing into the introduction of additional assistants. These assistance systems were aimed at supporting customers in successfully advertising the products on the platform. However, some of the retailers were using very diverse product hierarchies and metadata for their products. Therefore, analyses of a successful retailers had to be adjusted to fit the profile of other retailers.
Procedure
Initially, niologic brought together technical and operational stakeholders from product development in a workshop to develop the data strategy. Together they set the goal of a data-driven trading platform and trading support.
Profitable data was quickly identified; on the one hand, there were several databases and logging systems from which data could be merged. On the other hand, key figures and input parameters were identified, which would be available through a slight extension of the existing platform.
In the product development workshop, ideas for retail support were collected, evaluated and prioritized on the basis of data usage and opportunities.
As a first measure, various measures for location-based marketing were identified. Subsequently, a scoring algorithm was designed, which should give the retailer recommendations for action based on successful other retailers and best practices.
In addition, a generalized structure for a product hierarchy was designed with the customer to help evaluate similar products as the same for different merchants. Furthermore, the similarity of product characteristics should be used as a characteristic.
In a further step, niologic supported each development team in the expert interviews with new data scientists and in the idea of an embedded data scientist (see Product Analyst as part of the Scrum team).
Result and customer value
In the initial workshop, many of the client’s concerns regarding Big Data and AI could be resolved. Thanks to a data strategy, niologic could establish a tangible expansion of products. This was made possible by strong initiative from product management and an overall focus on product development. Furthermore, unnecessary obstacles in data editing could be avoided entirely, by introducing a smart data process. Later on, along with the client we conducted interviews to integrate one Data Scientist in each development team. In the meantime, the client was already able to implement several features of the roadmap, including location based marketing.Therefore, the platform could be successfully enhanced with mobile features and omni-channel marketing.
Thanks to this process, it was not necessary to create an entire data science department, so that our customer was able to save extra costs.
