Introducing data design

Design Thinking methods can be used to find the underlying motivations, feelings and core needs of the user, innovating new solutions and optimizing the overall user experience. Design Thinking methods can be applied also for data solutions, as they are designed most often with an employer, team or customer in mind. Focusing more time on building user understanding in the first phases of the data development project will save a lot of development effort and iterations in the later phases. It will also help to improve the collaboration and communication of the data solutions between developers, analysts and business users resulting in higher user satisfaction and ability to create value from data.

However, most often data solutions can't be designed only based on what the end user wants. The data might not be available or fit for the purpose. There might not be enough development capacity as there might be other solutions that have higher priority. The solution might increase the user satisfaction, but the wider business value creation potential might not be that high. There might be a data security or privacy risk  that the organization is not willing to take. Do these challenges sound familiar?

I have experience from utilizing these methods successfully in designing data solutions for media services while working for Sanoma Media Finland. In this article I interview Minna Kärhä, Head of Data at Finnair and Clemens Westrup, Head of Data Science at Sanoma Media Finland. They are both strong evangelists of this method and can share some benefits and concrete examples of using these methods in practice.


You can also check our recorded presentation of the learnings and experiences from Sanoma Media Finland and data design. The video was originally presented at the Data Innovation Summit 2020.



applying data design to designing any data solution

When developing the methodology, we often found us wondering what do we actually mean with a data solution or where this method can be applied. These are a few examples:

  • Dashboards and reports
  • Presentations with storytelling through graphs and insights
  • Data catalogs and dictionaries
  • Data and information architecture and tools
  • Algorithmic features in operational systems or digital services

We have split these into three categories:

Data design application areas

Designing data visualizations and user interfaces can be compared to designing web sites or applications.  In modern data visualization tools you can improve the data usability a lot by focusing on layout, navigation, buttons, use of colors, fonts and graphics and careful selection of data. The most important part in designing a data visualization solution is to understand the core user needs for the data and in which kind of situations the data is used. 

The data design method can be also applied to designing data tools and systems, by referring to data catalogs, application interfaces, data warehousing solutions and the overall architecture. The target of these kind of data solutions is to improve data accessibility, quality and usability in an organization.

Finally, data design can be also applied to designing and innovating new data services, features or products. This refers to for example recommendation engines, personalization features, chat bots, robotics and process automation and basically every feature in an operative system or process that uses processed data to perform better. This requires both innovation capability and understanding of what data is available and how it can be used for creating new features.


data design method in practice

It is hard to define a simple toolkit of methods that would work for every organization or project. We promote flexibility in selecting the methods, based on organization culture, team competencies and working models and solution complexity. The main principle in user-centric development of a data solution is to continuously improve your understanding of the core user needs and challeng your ideas about the potential solutions. Sometimes a coffee break or Teams discussion with a few potential users  might help you get on the right track. Sometimes data-based analysis on the current solution's user statistics is needed to understand the background a bit better. If the solution has a wide audience, you might want to try a survey to the users. There are also cases that are so difficult to get into that organizing a bit more structured design sprint or a innovation workshop is needed. If you know that it requires a lot of work to develop the solution, you might want to build a prototype first to check that you are going to the right direction with your plan. 

There is not a single set of methods, tools, interview questions or project plans that will work in every case. Some cases are very simple and straightforward, some need more investigation, analysis and design work. In general, typical design projects contain these type of phases: 

  • Discovering  and understanding the user needs
  • Evaluating ideas based on business value creation potential, technical feasibility and risks
  • Innovating potential solutions for solving the user needs
  • Experimenting alternative solutions and validating value creation potential
  • Productizing, testing, optimizing  and monitoring the final solution after it has been taken into use

why design thinking methods are not enough

When designing data solutions, it is quite unlikely that you can deliver exactly what the users need. Usually, there are a few things to consider, such as

  • What benefits would the solution create for the business, customers, employees or operations?
  • What is the technical readiness to start developing the solution, by meaning data architecture, accessibility, quality and fit to the purpose?
  • What is the operational readiness to take it into use, by meaning processes, competencies and willingness to adopt new tools?
  • Are there any risks of failing or delaying in the product development or does the data solution contain otherwise privacy or security risks?
  • Do you have enough development and maintenance capacity and skills available?

Since there are a lot of aspects to investigate and take into consideration, we decided to develop a canvas to help the design work. The canvas is applicable to any data solution design project. You can use it in a very light way to evaluate your existing idea or have a bit more extensive project by gathering information more widely across the organization. You can use it to evaluate a single idea, or generate multiple ideas for comparison. If you are interested on using the canvas in your work, drop us a note and we will guide you through it and provide you with a template. The canvas 1.0. version below is an example but when training the method to our customers, we usually make modifications to it and make sure that it fits to the overall design and demand practices of the organization. The canvas can be also embedded to other project planning and documentation tools, such as Confluence.

Data Design 1.0. Summary canvas guide


benefits in using data design

Minna Kärhä, Head of Data at Finnair, has been actively involved in developing the data design method. Minna shares her experiences on the importance of user-centric design in developing data solutions:

  • Crystallizing the right problem to be solved
  • Helping the team to make decisions and prioritize their work for the most value adding and best time-to-market solution
  • Ensuring the team will consider the critical aspects before finalizing the solution design
  • Creating great team spirit for success by empowering all the experts to formulate the final solution together
  • Increasing the overall understanding of data ownership and responsibilities and how data in created and used in the business

Minna Kärhä


experiences in utilizing data design

Clemens Westrup, Head of Data Science at Sanoma Media Finland, shares Minna's experiences. Along with his study background in Cognitive Computer Science and Machine Learning, Clemens has extensive experience as a Service Design practitioner. An alumni of the Stanford University based global Design Thinking program ME310, he spent two years teaching design methods and processes at Aalto Design Factory, as well as running trainings for companies. He is one of the originators of the methods in this post – tailoring user-centric design to data product development – and we spent a few years together iterating and improving the approaches in Sanoma Media Finland’s Data teams.

“It was always obvious to me that designing products, which rely on data as their main driver for value creation, would hugely benefit from user-centric methods, just as any other product or service does. But it is not exactly straightforward to apply for example classic Design Thinking, which has its roots in building physical products, in this domain. A seasoned service designer can easily adapt the process and methods, but what if you came out of your first workshop on user-centric design, ready to throw your new skills onto real world problems, but don’t know where to start? Therefore an opinionated but specifically tailored minimal framework can really help, for newcomers to design methods or experienced designers that are new to the challenges of data products."

Clemens continues explaining the most recent experiences on applying Design Thinking to Data Science: "At Sanoma we are taking an iterative grass-roots approach to bringing about the gradual adaption of user-centricity in our data teams and beyond. We organized a workshop together with our friends at Aalto Design Factory for the initial spark and to give familiarity. We then ran so-called Service Design Essentials projects on personalization in a 6-workshop format, in collaboration with Ruutu and Oikotie Työpaikat, to reiterate the process and tie it into the context of real product delivery with cross-functional teams. We started injecting specific tools, such as design briefs, into our planning processes and started encouraging experimentation with the tools in hackathons and ongoing projects. And most importantly, we are adapting it all to Sanoma’s culture and language. Because if you look closely, many people are already using these methods and processes, they might just not call them Service Design – and who cares really what you call it. It is our task to support these bottom-up drivers to foster wider adaption and exchange towards a more data-driven culture of innovation that we make our own.”

Clemens westrup

what is unique in data design method

 As Clemens mentions, it is not about what specific methods are used but more about how to adapt and tailor the methods to fit to the unique organization culture and environment. We also believe that data development projects have a few unique features that encouraged us to start developing the data design method. Taking into account the data availablility and quality, compliancy and usage restrictions and fit to the purpose are a few examples. We are also focusing on ensuring that the method stays practical and can be carried out  as a natural and easy step in the data development process. 

How the data design method differs from a traditional design thinking:

  • Specifically tailored and hand-picked methods for data and analytics development work instead of standard tools and templates
  • Practical examples and simple tips to be taken into use immediately without too many theoretical frameworks
  • Focus on doing and practicing with real-life cases 
  • Possibility for peer-to-peer coaching with other companies



contact us for help

We can offer wide variety of trainings to help you get started. One very effective way is also picking up a tricky user need or data solution idea and starting to practice the method with a real-life case. Please contact us for more information and help.

Contact details:

Iiris Lahti | email | Phone +358405188207