Bristol Myers Squibb

Biologic drug manufacturing greatly increases risk across a pharmaceutical site in areas such as quality, manufacturing and environmental health and safety. How might we forecast and mitigate risk with experience design and data science?

Bristol Myers Squibb
Design Lead


We know that risk on site can manifest in a variety of forms and locations. Our ambitious long-term north star vision is to empower manufacturing leaders to shift from remedying risk(reactive) to preventing risk (proactive).

Risks can also pose a threat to an organization's compliance in the eyes of a regulatory body, which implicates brand reputation.

The vision is to reinforce effort across global GMP standards and ensure the highest possible quality and safety of drugs for patients. The end state of 360° risk detection is about providing holistic insights to leaders about a site’s risk profile, traversing datasources and teams across sites.

Our ambition is to close the gap for site and network errors by giving visibility into real-time risk indicators.


The Risk Radar concept seeks to empower manufacturing leaders to proactively mitigate the potential manifestation of risks before they occur through:

  • The trending of site data sources over time to spot potential patterns which may indicate site conditions whereby risks could emerge.
  • Providing timely information on these trends to manufacturing leaders through an intuitive interface to motivate the creation of real-world mitigating actions.
  • The tracking of risk mitigation actions in one manageable central location.

This can be achieved by working around incompatibilities of data, enabling contribution and repatriating data to users. By constructing from individual monitoring contributions and providing sophisticated data analysis tools to strengthen regional capacity in coral reef monitoring, we can change the way in which we monitor coral reefs.

Experimentation Approach

Sprint approach

The team progressed our selected use case through a series of experiments during a seven-week sprint cycle. These weredeveloped to understand the concept’s desirability, feasibility &viability.

Scientist collaboration

A collaborative approach drew on the expertise of BMS subject matter experts (SMEs) who have experience of reactingand managing risk for their teams on the ground as well as those familiar with site data sets used for analysis.

Bristol Myers Squibb

User experience future concept


Design experiments were conducted including a user flow concept and a high - fidelity prototype. These were basedon user personas and scenarios identified in research. 


We believe a user would benefit from an experience instigated by a notified forecasted risk. The user will make a better informed decision with data insights to analyse, mitigate and manage the risk event before it manifests. 


SME and end user participants for feedback and validation. 

Success metrics

Usability and desirability scoring from SMEs and end users through scenario testing and survey.  


Positive validation of a user experience future concept that is usable and desirable with opportunity for expansion in future explorative work. Please, refer to the high – fidelity prototype and the concept flow for more detail.

User personas

Network level

Network level allows for the comparison of two or more sites to assess the firm’s risk from a global level.

Site level

Site Level refers to the holisticrisk assessment view of a site, including the aggregation of all work areas.This level also allows to explore within all work areas.

Department level

Department level refers to thehead of a department work area within a site. Examples include Quality,Manufacturing and Environmental Health & Safety.

Floor level

Manufacturing and Lab supervisorsat floor level will have permissions to view risks in the risk register butwill have restricted permissions on managing risk actions.

System mapping

We created a system map of the future state Risk Radar concept to better understand dependencies from a wider holistic view.

The users of the system are alerted of forecasted risk through the user interface, where they can assess, mitigate and manage risk.

Users’ tasks or jobs to be done are represented as features that are modular and can be added to as the scope of the application increases.

Features displayed through the UI utilize data identified in the risk landscape, in the case of our experimentation, this relates to talent experience and qualitydata.

Data will aggregate into the Risk Radar model to leveraged by data modelling techniques. Once a forecasted risk has been identified by the system it willnotify the users, completing the feedback loop.

Concept flow

User & system

The concept flow helped us to define the interdependencies between the user and the system and the definition of high level user actions and UI screens.

  • Data collection. Data is aggregated from manufacturing site systems including Workday, Infinity, LIMS, Veeva, EH&S and batch information.
  • Analytics forecasting. The system forecasts a pattern to identify conditions where a risk might manifest by leveraging time series modelling and will highlight potential risks to the relevant users.
  • Risk notification. The user is made aware of the risk through a notification with a time prompt. The notification will give some high level information into the risk area.Forecasted risks are collated in the Risk Register.
  • Risk analysis. A risk historical trend / forecasted trend can indicate a risk score over time and highlight potential future trends. On the analysis screen variables might be included with that risk.
  • Risk mitigation. The user can make an assisted decision based on insights from data modelling, combined with their subject matter knowledge to make a decision on risk mitigation.
  • Risk management. Project management screens of the risk backlog with information on risk responsibility and time to close out.

Validation & next steps

Scenario Testing  

The user experience was tested for usability and desirability with various stakeholders and end-users across Quality, Manufacturing, Auditing, Regulatory Compliance, Business Analytics and Insights. Data science outputs were referenced and reflected in the front end design. The concept and prototype was validated and a new phase has been established to further the R&D of the project with a view to industrialisation in the future phase.

Bristol Myers Squibb