Team members:
Michael Dickmann, Associate Director - Technical Product Management, AI/ML Foundations, with:
- Asheesh Chhabra
- Manoj Vig
- Jiri Janous
- Galina Spivak
- Alexandra Filip
- Vincent Capodanno
- Zdenek Kabat
- Richard Dobis
- Dharma Subramanian
- Mahendra Vemula
- Ajay Kumar
- Anupam Agnihotri
- Yasin Cagri Yigiter
- Samir Badhe
- Justina Ivanauskaite
- Ivan Merta
- Rudolf Kreibich
- Rastislav Matus
- Tomas Kremel
Country: United States
Organization: Merck
At Merck, known as MSD outside of the United States and Canada, we are unified around our purpose: We use the power of leading-edge science to save and improve lives around the world. For more than 130 years, we have brought hope to humanity through the development of important medicines and vaccines. We aspire to be the premier research-intensive biopharmaceutical company in the world – and today, we are at the forefront of research to deliver innovative health solutions that advance the prevention and treatment of diseases in people and animals. We foster a diverse and inclusive global workforce and operate responsibly every day to enable a safe, sustainable and healthy future for all people and communities. For more information, visit www.merck.com and connect with us on Twitter, Facebook, Instagram, YouTube and LinkedIn.
Awards Categories:
- Best Data Democratization Program
- Best ROI Story
Business Challenge:
Merck is market leader in biopharmaceutical research, manufacturing and supply. Collective efforts from various departments, regions and teams is pivotal to maintain the quality and accuracy. While digitalization and continuous medical advances are helping to meet the higher expectations and needs, it is also generating more and more data. Democratization of this data at enterprise level is a transformative business solution that will break down data silos, promote further collaboration, and empower employees with data-driven insights. Going "Enterprise" with data democratization makes the challenge multifold due to the associated cross functional expectations, policies and scale.
Merck explored various product options and most of them posed yet another barrier of "stiff learning curve". Irrespective of user personas they are needed to learn integrations and coding languages like Python, Scala, R, etc.
So the critical business challenge Merck contemplating at was "How to enable people, process and technology at scale and speed to be able to achieve the Enterprise level Data Democratization and Advanced Analytics goals".
Business Solution:
Four key components were identified to be successful in this initiative:
- Data Accessibility (Tech): Ability to integrate relevant data from various sources across the organization with minimal dependencies on tech teams
- Self-Service Analytics (Tech): Empower employees with the ability to explore, analyze and visualize data on their own, with no/minimal technical skills
- Data Governance Framework (Process): Establish guidelines, policies and procedures for data management to ensure quality, security, and compliance
- Training and Enablement (People): Comprehensive training and enablement programs to onboard employees with the necessary information, skills and support
Dataiku's low-code/no-code capabilities from data ingestions to advance analytics serving makes it a primary tool of choice, especially considering the barriers and associated key components identified above.
As a holistic approach towards enablement of Data and Analytics Democratization we decided to form three major workstreams:
- Adoption: Focused towards People, Process and delivery enablement
- Product & Engineering: Focused towards Product, Technology and controls
- Operations Support: Focused towards coordinated Technology and operations support
We started with an MVP approach with "Advanced Data Analytics Platform" and then building on Data and Modeling Capabilities, Industrialization Processes and User Enablement.
A learning path was devised to improve on overall Data Science literacy using Dataiku is designed with SME help from Dataiku Academy. This includes train the trainer, Certification drives, hands-on workshops, Webinars and similar periodic events. Followed by a simple User and Project onboarding experience.
With a continuous feedback loop with users a "Fit to Purpose" feature list is processed by our engineering team with Dataiku's Engineering SMEs. This enabled early adopters to realize their projects with no to minimal technology blockers.
An enterprise marketing and divisional engagement initiative with the help from our Dataiku SMEs also helped us to move from "Establish to Expand" mode by converting prospective ideas and opportunities into fully industrialized projects.
In brief, a comprehensive approach to include all aspects of enablement, execution and operations made it possible to hit the road and go for enterprise-level data democratization.
Business Area Enhanced: Analytics
Use Case Stage: In Production
Value Generated:
1. High Productivity Gain:
Even during our initial phase of "Establish", the Dataiku Platform users are already reporting an average 39.3% productivity gain. With further efforts on comprehensive adoption and structured platform services we are aiming to reach 60%+ in upcoming phases.
2. Extensive ability to do activities which were not possible earlier:
3. Enhanced Decision-Making: By enabling a wider range of employees to access and analyze data, promoting a culture of informed decision-making and driving better outcomes.
4. Increased Collaboration: With better access to relevant data, employees from different departments are in position to collaborate, share insights, and work together enhancing problem-solving and speed.
5. Business Agility: Business users can quickly explore and interpret data on their own, with no/minimal interventions from IT resulting in improved agility and responsiveness to evolving needs and market trends.
Value Brought by Dataiku:
Dataiku as a low-code/no-code MLOPs (Machine Learning Operations) technology has already started bringing in value of hyper productivity and agility as quick prototyping, test, iterate and industrialization of projects. These projects include wide range of use-cases in the areas of forecasting, Natural language processing (NLP), Computer vision etc.
Data pipelines are getting benefitted from Dataiku's ability to easily and securely interact with various enterprise data sources "on-premise" as well as "on-cloud". Complex projects are benefitted by the collaborative visual interface which enables Data Science teams to work hand-in-hand with the business on building and editing workflows.
In addition to this, in our case of "Enterprise Data Science Platform", where the need was more than just technology aspect of it. It was imperative to have extend efforts on user Trainings, Awareness drives (Internal Marketing), Adoption support, Divisional and regional level engagements and associated demand tracking, controlling, reporting. Dataiku's documentation, academy and Enterprise adoption support has helped achieve these goals.
Value Type:
- Improve customer/employee satisfaction
- Save time
Value Range: Hundreds of thousands of $
