Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
Search in posts
Search in pages

Productization of Data

Institutionalization of data products can help organizations scale their businesses.
  • Innotech
  • |
  • August 17, 2021

Productization of data is the process of treating data as a product; more specifically, the process involves translating insights gained from exploratory analysis into scalable models that can power data products. As such, productizing data can help integrate data science across all enterprise products and normalize its applications by making it accessible to even non-technical business users as clients can plug into a hosted cloud to access sophisticated data. This will be particularly useful for Bhutan considering the apprehensive nature towards technology adoption. Data productization can also increase business value through the use of data products to boost sales, deliver personalized experiences and improve products. 

One of the biggest advantages of productizing data science is the utilization of scarce data resources to carry out niche data science work, freeing companies from mundane or repetitive tasks. 

The other advantage of productizing data science is scale. Institutionalization of data products can help organizations take advantage of scale. For example, market mix models that are built for general merchandising can be leveraged for online grocery shopping and across different geographies if turned into a product. Such institutionalization of products can enhance efficiency and reduce redundancy. Data science productization can generate cost savings for the organizations by reducing duplication of efforts and facilitating reuse by using machine learning models as well as production systems for diversified applications. 

For Bhutan, data productization will be seen with the Integrated Data Platform project which will leverage data to gain insights that can be used for the creation of effective solutions and applications. 


  • Risk of losing data science resource investments.
  • Concept drift. Machine learning algorithms not connected to the constant feed of new data may make predictions that are less useful and accurate with the passage of time, resulting in lower accuracy of the statistical properties of the target variable that the model could be trying to predict. 
  • Risk of creating dysfunctional teams. we need to have both software engineers as well as data science experts in the deployment of machine-learning products with a strong collaboration between the two parties. 
  • Risk of unintended consequences. Risk of data bias, leading to unintended consequences in applications. 
  • Data governance issues (data privacy and security).
English (UK)