“Data platforms have two possible directions: data mesh or centralized Data Platform in terms of an operating model. Technologies are all developing very fast, and it’s getting more important to have internal competencies to lead the architecture.”, says in the interview for Hyperight.com Kristiina Tiilas, Head of digital platform at Outokumpu.
She is one of the speakers at the Data 2030 Summit this year. Before this event, Tiilas shares with our audience the journey of Outokumpu of creating a Data Platform that is compiled for their business users.
Find out the technical infrastructure of this platform, what methods and approach the company used in building it and what are some of the lessons learnt.
Hyperight: Can you please tell us more about you? What are your professional background and current working focus?
Kristiina Tiilas: I’m responsible for all data and analytics related topics within Outokumpu, from integrations to data platforms, advanced analytics and reporting. I have built our digital strategy, which I have done already in two other listed companies before. My academic background is in nuclear physics and I have a PhD in nuclear engineering.
Hyperight: During this year’s Data 2030 Summit, you will share more on “Making Data Platform compelling to business users”. What can the delegates at the event expect from your presentation?
Kristiina Tiilas: I always try to bring the business view to data, so my presentation brings up the digital maturity, atmosphere in the company and of course the way how data platforms can be made interesting and relevant for business users.
Hyperight: Can you give a brief background on your Data Platform journey so far? When did you start with it, for which purpose was it built, how far are you in the journey, and what does the tech stack look like?
Kristiina Tiilas: Outokumpu started the Data Platform journey many years ago and the main collaboration was with Microsoft. We focused the efforts on digital manufacturing and started with maintenance and quality use cases. The tech stack is mainly Microsoft Azure, but we also use Python, TensorFlow and others for ML use cases. Big part of our data also is located at SAP and Oracle platforms and our data strategy is under preparation as we speak.
Hyperight: Your presentation addresses one of the most important parts of any product, platform or service design: its usability and business or human-centricity. Many platforms, although perfectly built, can have a short life span due to integration challenges or being built without business consideration. You are addressing the importance of considering the user perspective when designing the platform to secure its usability and keep the interest level high and constant from the business and other users. Tell us more about it, and why do you think it is so important to take this approach?
Kristiina Tiilas: If the Data Platform is not taken in use by business users, it doesn’t really bring value to the company. The value comes from better decision making, process optimization and from new services that we can create to our customers. All these require changes in how people do their work, and not only IT/tech experts, but also customer services, sales experts, maintenance experts, process owners, etc. This is why the user-centric approach from the beginning to the end is key to success. The tools we create on top of a Data Platform need to be exactly correct in solving the selected problems and they have to be easy to use, or they are not taken in everyday use. Design thinking is the approach we use when building tools and services.
Hyperight: You will talk about turning to Design Thinking and Agile working methods to increase the probability of success of value realisation. How important were those when building your organisation’s Data Platform, and how were these methods used?
Kristiina Tiilas: Design thinking brings the users or customers to work closely with developers from the beginning all the way to the end of the development cycle. This way we can be sure that we are solving the right problem and that the user interface is created in the way the users want to see it. Agile development methods are typical for all tech development nowadays, we use SAFE for all our data platforms and take it in use also in traditional IT development now step by step.
Hyperight: You say that maturity growth, basic education and close collaboration through design thinking make Data Platforms valuable. What do you mean by that?
Kristiina Tiilas: Organizations maturity level defines how innovations are born, and how wide part of the organization is thinking about the possibilities that data and analytics bring. It’s very important for a wide part of the organization to understand the potential and think about the relevant areas where data could be used better. I have created training programs for Fortum and Kemira before and I’m creating an introductory program at the moment for Outokumpu. I want everybody to understand the basics: What is data and what kind of data do we have? What kind of analytics use cases can we create? What ML or AI means and when they should be considered, when it’s enough to visualize the data? What is our digital strategy and what are we aiming for? How can you get started with your own idea?
Hyperight: Can you share a bit about the success of the approach? How is the platform used today in the company, and how do you secure increased use?
Kristiina Tiilas: We use the Data Platform in many areas, e.g. in manufacturing, finance and sales. The growth of usage comes through use cases. We select the use cases we build and while building them we also build the capabilities and data sets for future needs. Step by step we have more data sets available for future use cases. We also celebrate successes and share experiences across the organization.
Hyperight: What are your lessons learned from the journey so far, or what would be your recommendations to those just starting building a Data Platform or in the middle of it but struggling with user interest? Where should they start, and what to pay attention to?
Kristiina Tiilas: Start with use cases and build the Data Platform capabilities through real cases. Plan the platform capabilities at the same time and have one person responsible for the Data Platform as product owner. Build core competences in-house and use externals as needed. Select the partners well, the biggest ones are not always the best ones for you.
Hyperight: From your perspective, how do you see the future of Data Platforms? Any trends you see more of in the upcoming 1-2 years?
Kristiina Tiilas: Data platforms have two possible directions: data mesh or centralized Data Platform in terms of operating model. Technologies are all developing very fast and it’s getting more important to have internal competences to lead the architecture. Agile ways of working are also taking a bigger role in both advanced technologies as well as traditional IT systems. The collaboration model and clear data strategy is still missing in many companies.