F. Hoffmann La Roche AG - Running a Launchpad Program to Expand the Organizational Knowledge and Utility of Data

Name:

Omar Khawaja
Riaan Badenhorst
Dr. Frank Block
Ines Dias
Mark Bucherl
Nakul Aggarwal
Dr. Mohammadjavad Faraji

Country: Switzerland

Organization: F. Hoffmann La Roche AG

Roche is a global pioneer in pharmaceuticals and diagnostics focused on advancing science to improve people’s lives. The combined strengths of pharmaceuticals and diagnostics under one roof have made Roche the leader in personalized healthcare – a strategy that aims to fit the right treatment to each patient in the best way possible.

Founded in 1896, Roche is the world’s largest biotech company, with truly differentiated medicines in oncology, immunology, infectious diseases, ophthalmology, and diseases of the central nervous system. Roche is also the world leader in in-vitro diagnostics and tissue-based cancer diagnostics, and a frontrunner in diabetes management.

Awards Categories:

  • Most Extraordinary AI Maker(s)

 

Business Challenge:

As part of a 10-year plan, Roche has greatly invested in digital transformation to garner the power of data and analytics in generating actionable insights from medically relevant and business-related data as a key competitive advantage in the healthcare industry. Driving personalized healthcare, digital diagnostics solutions, and operational excellence are just a few examples of using data and analytics at Roche to foster healthcare technology.

Roche Data Science teams help Diagnostics and Pharmaceutical business domains to become more data-driven, innovative, and efficient in delivering insights for smarter decision-making, faster prototyping, and products that leverage data and advanced analytics (AA). In 2017, a handful of Roche data scientists began to explore and explain the potential of applied advanced data analytics to business and operations practice leaders. Since then, the number of advanced data analysts and data scientists has grown exponentially across Roche. Now, many advanced analytics teams routinely work with practice leaders to quickly take theoretical proof-of-value (PoV) use cases and move them into live production services and solutions that propel end users’ data-driven objectives.

Expanding the organizational knowledge and utility of data across functions requires tools that harness coding potential for data scientists (today mostly Python and R), while also being accessible to non-technical decision makers. The collaboration among data analytics professionals and business users who own domain knowledge and expertise, and meeting the needs of other data science personas, are all key to the success of scaling up Machine Learning (ML) and Artificial Intelligence (AI) use across organizations.

At Roche, we searched for an enterprise-scale data science workbench that was scalable, supported various use cases, and accelerated project time to value by reducing the PoV time as well as the time required to industrialize or productionize AA use cases, thus enabling a robust MLOps model. Furthermore, we sought a more inclusive AI framework to address the needs of various data science personas across Roche to unlock the value behind data and analytics for multiple functions within Roche and its affiliates.

 

Business Solution:

The Diagnostics Informatics Data Science team first carried out an evaluation in 2021 to identify an adequate Data Science Workbench for its daily business. Aside from technical capabilities, an important requirement was to find a tool that would enable various data science personas and was also attractive to less-technical, business-oriented collaborators.

At the end of the evaluation, the Dataiku Data Science Studio (DSS) was shortlisted for the following reasons:

  • User-friendly to “non-coders.”
  • Availability of a broad spectrum of AI / ML / AA techniques, models, and methods.
  • Streamlining the productization process on ONE platform.
  • Possibility of using R and Python to extend and customize the functionality offered by the tool.
  • Cloud-based and scalable.
  • Good architectural fit in the existing data stack in place at Roche.

The Diagnostics Informatics Data Science team started using Dataiku DSS for daily operations in a pilot program that began in January 2022 and will last until the end of 2022. This data science team was trained and used Dataiku for PoV initiatives, collaborating directly with business subject matter experts.

Following the kick-off of the pilot program in Diagnostics Informatics, an opportunity to offer a Dataiku trial to all interested Roche employees was requested to evaluate Dataiku DSS usage for their data-driven and AI activities. In April 2022, we initiated a six-month global initiative called the “Dataiku Launchpad Program.” The aim of the Launchpad program was to let various data science teams across Roche have hands-on experience using Dataiku DSS to support or challenge the evaluation carried out in 2021. The Launchpad program would validate the research and collect enough data points to make an informed decision for the future of Dataiku at Roche.

The program’s objectives were: (i) to prove Dataiku’s ability to onboard and enable hundreds of users in a short time frame, and (ii) to allow stakeholders across Roche to evaluate Dataiku as a workbench for their own uses. Throughout the program, regular enablement sessions addressed ideation, coaching, success sharing, and focused development. Question and answer sessions were also organized to support participants' exploration of the DSS platform and assess whether they would like to adopt the tool in the future.

Business area enhanced: Accounting/Finance/Analytics/Internal Operations/Manufacturing/Marketing/Sales/Customer Relationship Management/Supply-chain/Supplier Management/Service Delivery/Disease Analysis (R&D)

Use case stage: Proof of Concept/In Progress/In Production

 

Value Generated:

In the first three months of the pilot program, 30 data scientists took training on Dataiku, and 10 PoVs were created. This pilot validated the research hypothesis that Dataiku could reduce the start-to-finish time required for PoVs.

Passing through the fifth month and entering the final month of the Launchpad Program, as of September 2022, we observed a high level of interest and engagement from multiple analytical teams, who expressed enthusiasm for working with Dataiku DSS in the future. The following are engagement statistics that show the usage of Dataiku DSS in Roche for the pilot and Launchpad programs combined:

  • 200+ program participants from many regions, functions, and roles.
  • 150+ DSS users.
  • 120+ Roche colleagues trained.
  • 520+ courses taken, with an 80% completion rate.
  • 450+ DSS projects created (including training).
  • 150+ business-related use cases created.
  • DSS usage has risen to 650+ hours a week.

Upon completion of the Launchpad Program in October, we will assess the desire of all participants to continue working with Dataiku DSS. Based on the final evaluation results, the future of Dataiku at Roche at the enterprise level will be determined. The true adoption of Dataiku DSS at Roche can only be confirmed after it has been extensively used at the enterprise level during the non-experimental phase.

 

Value Brought by Dataiku:

Technical capabilities and broad access were critical factors in the analytics platform evaluation and selection of Dataiku. While assessing several data science workbenches, it was the combination of a wide array of AI and ML functionalities, programming and scripting support, end-to-end ideation to productization support, and Dataiku’s accessibility for non-technical colleagues that most influenced the selection of this solution for the pilot and Launchpad programs.

The following summarizes the benefits that the Diagnostics Informatics Data Science team has realized during the pilot program:

  • Dataiku DSS helped us increase speed and agility in executing our analytics-driven initiatives.
  • DSS helped us attract Roche colleagues from beyond the data science community to the analytics space, thanks to its low-code and no-code features, as well as its user-friendly environment.
  • Dataiku dashboards helped us incorporate business feedback more regularly and efficiently than before by increasing collaboration among different teams.
  • The smooth process for deploying solutions – from design to completed automation node – significantly reduced production time, accelerating value generation for our customers.
  • DSS features such as scenarios and metrics made automated scheduling and monitoring of our live applications much easier than before.
  • The support from the Dataiku team during the pilot and Launchpad programs helped us educate our colleagues through numerous training, coaching, and Q&A sessions and enhanced consideration of using Dataiku DSS as a valuable addition to a team’s data analytics toolbox.
Share:
Version history
Publication date:
03-09-2023 12:39 AM
Version history
Last update:
‎09-15-2022 02:39 PM
Updated by:
Contributors