UNIFORM PRODUCT SPECIFICATIONS THANKS TO AUTOMATED ONBOARDING

With more than 270 specialist shops in 19 European countries, BAUHAUS holds a leading position in the workshop, home and garden sector. Craftsmen, tradesmen and private customers throughout Europe rely on a large selection and first-class quality at favourable prices, both in the specialist shops and in the online shop.

The general market trend of growing online business can also be observed at BAUHAUS. In order to take this development into account, the company has invested heavily in online trading.

One of the measures being pursued is an expansion of the product range available online. A decisive factor here is the efficient creation of new products in high quality that is customer and search engine friendly.

The DIY sector is characterised by an extensive range of products which have numerous different properties. To ensure the quality of the product data under these circumstances, BAUHAUS relies on a detailed product data model. Entering the data of new products manually is cost-inefficient, error-prone and not scalable in the long run. For this reason, BAUHAUS has opted for an automated onboarding solution.

HOW ONEDOT HELPED

BAUHAUS pursues a product group-based strategy to expand the online assortment. Matching suppliers are bundled per product group and all catalogues are uploaded together to the Onedot platform.

The structured Onedot onboarding process is then initiated. The categories assigned during the initial creation in the ERP are taken over directly by Onedot. This is done by matching the supplier articles with an ERP export. This allows category-specific attributes to be identified, which are assigned within the framework of the attribute mapping.

The next step is the attribute mapping. Here, the attributes of the suppliers are assigned to the BAUHAUS attributes. Based on statistical properties, word similarities and semantic methods, the Onedot software now automatically suggests mappings. The user can view these suggestions online in the Onedot user interface, validate them and adjust them if necessary. If no mapping suggestions could be determined, the user can instead check whether another mapping is possible.

The feedback received in this way is now fed back into the Onedot software, which updates the supplier profile as well as the machine learning models. In the next onboarding of this supplier, the mapping suggestions updated in this way are reused, resulting in even more accurate suggestions for new attributes in each run.

In the next step, normalisation, numbers and units are normalised and converted. For attributes that are based on a list of values, the mapping suggestions are again automated by the Onedot software, which can be viewed and adjusted by the user if necessary. This automated creation of suggestions saves a lot of time compared to the manual creation of mappings. Each suggestion of the software has a confidence level from low to high, which indicates how valid this mapping is. This helps the user to check only those suggestions for which a low confidence level has been achieved by the software.

Such automatically suggested mappings translate the supplier's data structure quickly and efficiently to the BAUHAUS product data structure. Once this data transformation is complete, an import file is created which BAUHAUS can import directly into its own PIM.

THE RESULT

BAUHAUS can bring products online much faster by using the Onedot platform. This means that it is now possible to flexibly expand the online range. The time-to-market can also be significantly reduced. Through the precise, category-specific attribute mapping via the Onedot software and the subsequent standardisation of the attribute values, BAUHAUS can meet its high requirements for product data quality.

This leads to a high-quality customer experience in the online shop and an increased purchase rate. Thanks to the structured output of product properties, customers can precisely search for individual products via search filters. In addition, the product data team is greatly relieved by the automation of repetitive, manual activities.