Design Sprint

Search applications are best developed in a series of sprints. A design sprint is basically the first sprint in this series; a week in which we go through all stages of a search design cycle together.

In a design sprint we get to know your users, your data and your problem, we investigate different solutions together with your most important stakeholders and build a first prototype that your users test.

Meanwhile we create the three interdependent parts of your initial search application: a knowledge graph containing your data, search algorithms that select and rank relevant data and a user interface that allows your users to access them. We apply design thinking to the creation of these parts:

Understand

Empathize

We start by talking to you (the product owner of the envisioned search application) and other stakeholders to get a good understanding of your organization, your requirements and your expectations. Next we talk to your users about the tasks they are engaged in, the information systems they currently use and the problems they encounter in doing so. And finally we interview data engineers about the datasets that exist, whether they are relevant for the search tasks at hand and how they can be made available.

Define

Subsequently we group and prioritize the collected search tasks, identify the concepts that make up your domain, establish which concepts your users need to fulfill each search task and what aspects of the concept (and any related concepts) determine their relevance. Finally we determine which search tasks to focus on in the remainder of the sprint.

Explore

Ideate

Next we start sketching the three parts based on what we've learned: we sketch a model of your domain that shows what constitutes the identified concepts and how they relate to each other, we write down initial specifications for all algorithms denoting the input, the output and a high-level description of their operation and we sketch wireframes for every interaction.

Prototype

We now turn the most promising ideas into concrete, tangible prototypes: we turn your data into an initial knowledge graph, graphically create search algorithms in Spinque Desk that determine which data is selected and how it is ranked and build a basic user interface.

Materialize

Test

Finally we align the three prototypes and turn to the users. We ask them to search for information they need to fulfill their usual work tasks using the combined prototype. During these tests the users assess the relevance and suitability of the search results, the way in which they are presented and the ease with which they are found.

Implement

At the end of the design sprint we evaluate what we have learned together and decide whether/how to proceed. We now know your users, your data and your search problem, have a shared understanding of the requirements for your application and are well able to map out a possible follow-up process.

In a design sprint we thus help you to clarify your search problem and to understand the information needs of your users. It allows you to make well-founded decisions on how to set up your project and enables you to fill a backlog with search tasks that your application should support. To top it off you can use your initial search application to further develop your ideas with users and stakeholders.