Meet the Team: creating tailored search

Posted on 03/11/2022 by Peter Tessel

We power search for clients in many domains: eCommerce, public administration, cultural heritage, business operations, etc; organizations with large amounts of heterogeneous data. Although the domains differ, our challenge is always the same: enable users to efficiently accomplish tasks that involve search. We define search broadly: not just finding a specific item or exploring a topic but also recommending items, personalizing results and analyzing data.

Search questions: who, what, which, where, when, why, how?

In this blog we introduce the three different search designers who as a team create tailored search that does efficiently support users. Spoiler: you are one of them!

The Case for Search Design

The figure below shows the setup of conventional search. Designers create the frontend and determine how users interact with the system; how they express their queries, process information and filter results. Engineers manage the backend and maintain the system; they determine how data is indexed and how the system is configured.

The setup of conventional search.

This conventional setup has several shortcomings. The first is caused by the way the data being searched is stored. Users often require data from different sources to meet their information needs. While most applications offer search functionality, they generally store their data in separate proprietary data sources. As a consequence users search datasources separately and have to combine their results, while they really need search results that combine all relevant data for them. In addition, users are often looking for information that is not explicitly represented in the data sources. Data takes on meaning in the applications that use them, but this knowledge is not always explicitly available in the sources that store the data. As a consequence users frequently have to compose a desired result from a number of partial results or have to search for a desired result in a more comprehensive result, while they really need search results that exactly match the concept they're looking for.

We solve these problems by converting data from different sources into a knowledge graph. A knowledge scientist extracts the data from all relevant sources and integrates it, adds knowledge about your domain and thus creates a conceptual layer above your existing systems.

Knowledge graph added to the conventional setup of search.

The second shortcoming of convential search is the way data is retrieved. Or more precisely: who determines the relevance of the search results? Conventional search requires people with specific knowledge of the search technology to manage and configure the system, so in practice programmers are usually in charge of relevance. As a consequence users are often provided with a catch-all algorithm, while they really need an algorithm that is tailored to their specific search task.

We solve this problem by enabling you to take on the role of information specialist and to create search algorithms that answer your queries yourself. After all, you know your domain and what your users are looking for best. Our application Spinque Desk enables you to combine building blocks that perform various operations on the knowledge graph and so to build search strategies that answer your domain specific queries perfectly.

Search strategy and knowledge graph added to the conventional setup of search.

You expose your search strategy via a dedicated API using Spinque Desk. We take care of performance, scalability and efficient execution. We thus put you in charge of your search by replacing the black box of conventional search with a knowledge graph, one or more search strategies and a matching API. If new requirements or needs arise you simply adjust your existing strategies or create new ones.

API, search strategy and knowledge graph added to the conventional setup of search.

Now that the knowledge scientist and information specialist have created tailored search, the interface is geared to it. Here the UX designer comes in. She uses the concepts from the knowledge graph to best match the expectations and needs of the users and so to optimally support them in performing their domain specific search tasks.

Spinque's approach to search.

Search applications are best created together

In a search design process the knowledge scientist, information specialist and UX designer thus jointly create tailored search that efficiently supports users in accomplishing their tasks.

The UX Designer, the information specialist and the knowledge scientist.

This is the first blog post in a series about the search design team. In three upcoming posts we will take a closer look at the work of the knowledge scientist, information specialist and UX designer. Stay tuned!