Development of an internal search engine

How does a company like Daimler centralize relevant knowledge and make it accessible to all employees?

To answer this question, we had the privilege of working with Daimler's CIR (Corporate Information Research) team for 18 months, together with our partner Portaltech Reply, to launch a new intranet application in the form of a document search engine.

The challenge

Daimler provides its employees with company-relevant knowledge in the form of articles, videos, books, statistics, etc. via a portal. However, as the number of documents grew rapidly over the years, the application was neither architecturally nor technically able to meet the new knowledge management requirements.

More than 50 different databases, APIs and information sources can only be accessed individually by employees. In total, more than 2 million documents were stored in a decentralized manner, resulting in a cumbersome research environment that was difficult to navigate.

Replacing this system and providing a centralized database with a single interface for staff requires answering several fundamental project questions:

• How can documents from disparate sources be migrated and aggregated into a single knowledge base?

• Which technologies can be used to implement the application in a high-performance and agile way, given the enormous amount of data?

• What is the right user experience concept?

• How do we enable nearly 300,000 employees to get the best possible search results based on their job descriptions and interests?

The joint project was based on a "greenfield approach", which allows for a concept without restrictions due to previous work or projects. As a result, the project team was able to develop a user experience concept that was methodologically, technically, and content-wise capable of delivering a promising application.

The php framework Symfony was used to manage and process the documents, allowing for flexible implementation of the project's individual requirements. Elasticsearch is used to build a high performance database, search index and search engine. The world's leading open source search engine stores data records in the NoSQL format and can thus offer a high-performance output of the content in the frontend via its REST service.

The implementation of the frontend using the JavaScript framework Vue.js allows the development of individual and fast search interfaces with declarative components to build the structure of the application according to own requirements.

The technical implementation of the application was realized in collaboration with Portaltech Reply and organized according to the agile project management method "Scrum".

User Experience

A good UX design process starts with elicitation of business goals, customer interviews, requirements definition, prototyping, and regular user surveys. The focus is on designing the application to meet the needs of the users. This is the only way to achieve the highest possible acceptance and usage by the target group.

The first step in the cycle is to define key users. These so-called personas form the basis for the definition of use cases and map the concrete needs of the users.

These use cases are outlined in wireframes (rough visualization of layouts for content structuring and hierarchy) and improved with the team and test persons in the form of surveys & interviews.

The insights gained form the basis for the interface designer to develop concrete UI components and create designs.

According to Scrum, design and technical implementation were done step by step. The result of the Scrum methodology is incremental software development. This allows the project team to test the first functions & features with the users on the "live product".

Especially for a search engine application, it is necessary to conduct user interviews on the technically realized product instead of on the click dummy. This is the only way to check whether the functionalities have been implemented intuitively and performantly enough for the user. Thanks to the close cooperation in the interdisciplinary team of UI/UX designers and developers, it was always possible to implement the best solution taking into account user feedback.

Thanks to the increased transparency and clarity for the project team, the application was implemented along the user needs.