Precise power: About flexible solutions in OBI's online shop search

Online store searches are becoming increasingly complex, which is why our client OBI was faced with a major challenge: the store was based on a search technology where the ranking of search results was only slightly customizable. We put the project in the spotlight.

Problem definition

• Domain-specific requirements (such as variable prices and stocks in the markets) had to be taken into account

• high configuration effort of the previous search

• Insufficient customizability of the search result ranking

Solution approach

• Development of a new search tool based on an open source search engine

• Deployment in the cloud with a Java backend and Vue frontend in a Kubernetes environment

• Indexing of the product data in the open source search engine

• Customized search architecture with flexible, scalable structure

• Integration of new features such as dimension search

Result

• Significant improvement in search and performance

• Higher conversion and click-through rate

• Cost reduction through cloud migration

• Reduction of the configuration effort

• Three times faster page speed

• Increase in turnover

• Preparation for internationalization and AI integration

With this in mind, our team of neuland experts stepped up to the plate with the aim of developing a customized solution together: A completely newly developed search tool, based on an open source search engine deployed in the cloud, coupled with a Java backend and a Vue frontend in a Kubernetes environment. This advanced architecture provides OBI with what was previously missing: flexibility, scalability and, above all, an easily searchable product database that adapts seamlessly to the company's individual needs. This enables a massive reduction in configuration effort and thus revolutionizes the search in our customer's store.

OBI - A distinctive branch and franchise system

We have been supporting OBI in the operation and ongoing development of its online store since 2010. With more than 640 stores in Europe, 349 of which are in Germany, and an online store, OBI is one of the largest DIY stores in Germany. Originally, the online store was conceived as an insight into the product catalog for customers. The initial assumption was that customers would research online and then go to the stores to buy - a concept known as ROPO (Research Online, Purchase Offline). In recent years, however, OBI has increasingly focused on the online area and expanded this together with neuland, among others. Online sales are becoming increasingly important, which is also influenced by the franchise structure of the stores. The stores are operated by various retailers as franchise partners who sell the OBI range locally.

neuland's answer to complex search requirements

The previous search technology in OBI's online store had several limitations as well as the challenge of taking into account the domain-specific requirements, such as the variable prices and stocks in the different stores.

The new architecture enables a flexible, modular structure that allows the search functions to be customized to OBI's needs. The project team manages the entire search vertical, which means that the team takes responsibility and builds up domain and technology knowledge. In the DevOps area, they develop and maintain the platform, manage all relevant data and independently develop new features from the front end to the back end.

Indexing the product data in an open source search engine makes it easy to adapt the search criteria, as the most common functions are already implemented with high performance. A significant improvement in the search is achieved through new features, such as the search for measurements, which delivers precise results even if the query and product data refer to different units of measurement.

In addition, the development time for new features is significantly faster, which means that complex requirements such as search engine optimization (SEO) and the integration of retail media in the store can be implemented. The team no longer has to implement workarounds to circumvent the limitations of a black box, but has the opportunity to develop and test its own solution concepts for every need or problem.

Measurable success: the impact of an optimized search

It is not only on a technical level that the importance of search is increasing compared to category navigation in the store; the effects of search are also clearly visible in the sales figures: an increase in performance, a higher conversion rate, a higher click-through rate and increased sales, as tests have shown. The new solution enables up to three times faster page speed (time-to-first-byte). Irrelevant search results are significantly reduced, as confirmed by qualified user testing.

In addition, the search developed in-house achieves a significant cost reduction compared to the old, license-based search technology, particularly through cloud migration. Our next steps will be internationalization and the integration of AI. An initial proof of concept (PoC) for vector search has already been successful, in which a similarity search was carried out on product embeddings.

The collaboration in our team proved to be very valuable: we delivered rapid development and implementation of new features, even during ongoing operations. Our client OBI's trust in our strong commitment to innovation creates the space to fully utilize our skills and expertise for tailor-made solutions. The development of a new, more efficient search not only illustrates the successful transformation on a technical level, it also provides a decisive competitive advantage in the world of e-commerce.