Data is the most valuable asset in growth-oriented companies. Data is like any other commodity and requires refining to gain more value in the eyes and processes of customers. Yet data products are not enough anymore. Centralized models based on integrating datasets to data lakes and data warehouses are challenged with more robust distributed border-crossing and servitized solutions. Thus we will face the same servitization in the data economy, that we have witnessed in the digital entertainment and aviation business. Data servitization combined with subscription economy models is already becoming the de facto approach in data monetization. 

Data-driven growing market 

Data-driven digitalization is proven to be profitable since according to a survey conducted by McKinsey Global Institute data-driven organizations are over 20 times more likely to acquire customers, half a dozen times as likely to retain their customers, and 19 times as likely to be profitable. Conservative estimate of the market size for data in the global hospitality industry alone was US $43.2 Billion in 2018, and it has been doubling its size every three years. According to DataLandscape in the EU the estimated size of the Data Market was 60 Billion euros in 2016. 

Value creation is driven by productizement and servitization 

Data monetization which includes data servitization is part of the digital transformation. The digital transformation leads to major changes in established value creation structures and traditional business models of companies. Data is increasingly used beyond the improvement of internal processes by serving as a strategic resource for the development of data-driven innovations and business models. This data-driven innovation and creation of economic value is less and less created by a single organization or in traditional value chains but instead takes place in cross-industry, socio-technical networks – so-called data ecosystems. Given the nature of data ecosystems requiring border crossing activities in value creation, data used in the process must be packaged into products and services for more efficient reuse and sales. Jedd et al state in Harvard Business Review that data should be approached with a product mindset. 

Service-Dominant logic success stories 

At the same time, servitization is a megatrend and we’ve all seen and experienced the change. Servitization has been defined in the research literature as “The transformational processes whereby a company shifts from a product-centric to a service-centric business model and logic.” The older generation (like me) still remember the time when we purchased movies by purchasing products – DVDs. Then suddenly that was obsolete and we got the movies from a service called Netflix.

Drawing 1: Examples of servitization are Netflix and Rolls Royce. Both disrupted their own industry with help of servitization.

What happened was that Netflix servitized the movie experience. Servitization is not limited only to digital business. In the old days of the goods-dominant period jet plane manufacturers bought the engines from Rolls Royce with optional maintenance services. Then Rolls Royce decided to servitize the engine business and started to sell flight hours. These two exemplify just the tip of the iceberg. Data is no exception and is already partially servitized.

Data will be servitized because of business needs 

The claim is that data servitization will drive the data economy. What happens to data products then? The movies as products did not disappear when we started to use Netflix. What changed was what we were paying for and the product became fully digital. We used to pay to own the physical DVD. Now we pay for the 24/7 access to the movies and tv series (digital products). We don’t need or want to own the movies since we can access the content, again and again, every day and every hour. Owning has become obsolete. We are witnessing the same phenomenon with data commodities now. 

We are now entering the data servitization period. The more traditional data product-driven paradigm was about productizing data, datasets, ownership, data integration, and customer pull. The new Data as a Service paradigm is about servitization, data streams, access and rights to data, subscription, and pushing changes to the customer. The rule of thumb can be summarized in two sentences. First, productize data for efficient internal reuse and sales as well as feeding the partner value chain. Secondly, servitize data for maximum revenue and direct value for customers.

From products to services 

We need to understand that some of our customers have different problems and some of those are solved with data products and some with data services. Business developers need to understand that data is more often productized to enable faster reuse and enable monetization via sales at the same time. Data must be at least productized to remove the need for tacit knowledge regarding the content and use. 

The current event-driven architectures push data providers to take the next leap forward and require data as streams – as a service. At the data provider side, data streams are servitized data products. 

Drawing 2: A simple example of data servitization via productizement. Note that servitized data has more formats than just a data-driven data stream. Other common formats are data stories and dashboards.

From integration to subscription 

Still, part of the data economy is built upon selling datasets. Datasets are static collections of data, often hundreds or thousands of lines or objects of data or even more. Those datasets are purchased from various data markets and alike. After that, the dataset is integrated to a bigger pool of data also known as data lakes and data warehouses. The data is then utilized by the data scientists who for example build visualizations from it for business decision-making purposes. Very often the reality is messier and according to The State of Data Science 2020 report by Anaconda data scientists spent 45% of their time cleansing and loading the data before they can use it. The above describes the traditional data integration approach in which the pure data product approach works still pretty nice. 

Yet a significant amount of data has value only for a moment. This is the case at least with IoT control data. Data is used as input to adjust systems and processes. Thus data must be delivered on-demand, be accurate, and available via an API – as a service. More often the customer is passive and receives the changed information from the data provider. In other words, the customer subscribes to the data stream and waits for new input, and then reacts to it in their own application or service. 

Pros and cons of data as a service 

Not all data streams or datasets will be real-time. We still need to build trends, predictions, and other charts in the future. For those purposes, historical data (datasets) are needed. Most likely your organization needs to adopt a hybrid strategy and enable use of both data as a product (dataset) and data as a service (streams and other formats). If your organization is going towards monetizing data, I recommend getting ready to servitize your data assets. Servitization of data is already happening and looks like it will take over product thinking soon.

Jarkko is the Country CDO Ambassador for Finland, Chief Data Officer at Vastuu Group. Prior to his employment at Vastuu Group, he served as the Chief Operations Officer at Platform of Trust. He is currently pursuing his 2nd doctor’s degree around the design driven data productizement process, which binds together data products as well as data services and data strategy in companies. Jarkko is also Data Economy Advisor at Techie Stories Ltd, cofounder of Data Product Business and creator of Data Product Toolkit.