Sanoma Media Finland (SMF) is the leading Finnish multi-channel media company whose products reach most of the Finns every day. The digital transformation has been rapid in media and Sanoma has been one of the leading companies in taking advantage of the development of the technology and the media consumption trends. Data is playing and continues to play a highly strategic role in the competitive position and performance of all media industry businesses.
At the Helsinki Data Science meetup at Sanoma on 6th November, it was therefore interesting to hear how Sanoma has succeeded in this transformation into a truly data-driven business with data democratization, real-time content personalization capabilities, productized and award-winning data services for advertisers, digital user experience optimization capabilities and revenue streams.
The evening started off with Iiris Lahti, Development director in Data Utilization from Sanoma Media Finland , and DAIN Studios’ CEO Ulla Kruhse-Lehtonen, walking down the memory lane and revisiting the journey of the team that Ulla put together before venturing off to co-found DAIN Studios in 2016. “One of the biggest changes I’ve seen in the past 6 years in the media industry, has been consumers willingness to pay for digital media services” Iiris summarizes. “Sanoma has also succeeded in the digital transformation of our product offerings and business models. The digital growth strategies in 2012 sounded back then like a stretch goal, whereas looking at the numbers now, those goals have already been surpassed.”
When growing a data-based business, Iiris recommends starting with building the trust towards data. “Tools and technologies are obviously important in managing high quality data platforms and services, but they are not enough. Companies need to also focus on building the competences to putting the data to work and finally growing the roots for a data driven culture. Today analysts and data scientists are an integral part of the digital product development and sales teams at Sanoma, and teams have learned how to collaborate and create value together”
Both Ulla and Iiris shared some tips for other organizations to follow if they want to become data driven.
Caption: Iiris Lahti and the key lessons on how to become a data driven organization.
Next up, Valtteri Vartiainen and Risto Tuomainen presented Sanoma Media Finland’s data products and services for their advertising customers. rs. Valtteri first introduced the data products, for example catalogued target groups available for all advertisers and tailor made target groups such as interest profiles or geographies. Sanoma also provides data partnership services, where the advertising customer can utilize custom segmentation and insight generation capabilities together with their own data, supporting their marketing decision making.
Risto Tuomainen, a data scientist, provided a case example of such collaboration with a retail customer. Comparing the data from a client company surveys to Sanoma’s panel, the model Risto portrayed was able to predict which client company customer segment the Sanoma readers belong to.
In the presentation that followed, Clemens Westrup and Risto Kivioja shared their experiences from testing Design Thinking and Lean Service Creation practices.
Clemens first talked about exploring the use of service design tools combined with data. He has been intrigued by service design as a methodology, as he sees it as a natural fit to the exploratory nature of data science work. “It’s a mindset for prototyping you need to have.”
Before one can start prototyping, the collaborators need to do some divergent thinking, that is, to produce a high quantity of ideas to widen your perspective while thinking less about their quality. This can be difficult for analysts who are solution driven and therefore usually in a convergent state of mind. “Usually your first idea is not your best – even from a statistical standpoint that seems highly unlikely. Great ideas also don’t happen in isolation, and it’s usually not even ‘your idea’ to start with, but the result of previous knowledge and ideation with others. Instead great ideas happen when people share their different perspectives, don’t hold back or judge, listen and build on each other’s suggestions”, Clemens explains.
Risto then shared the experiences around the first pilot with Ruutu, a video streaming service by Sanoma. Utilizing lean service creation principles, the team did a bunch of interviews, then workshopped on personas. A ‘dark horse’ prototype was also created: something totally different, and fun – in this case the option for the user to get a new and surprising set of recommendations. A version of that prototype was eventually introduced in the service.
Last but not least, data scientists Sergey Chernov and Max Koppatz presented real-life challenges in building personalization engines from scratch. The case in question was cold start problems in content based job recommendation on the Oikotie Työpaikat job listings platform, specifically focusing on how to match job postings in different languages. Sergey and Max proceeded to solve the problem in real time. They demonstrated how to apply Concept Net NumberBatch and lemmatization with prepared coding, that you can experience in this tweet by Iiris: