Increasing reliability from 92% to 96%!
GE / BHGE
Engineers managing offshore oil platforms face challenges like costly shutdowns and non-standardized safety processes. To address these issues, we developed Plant Operations Advisor (POA), a product that enhances efficiency, reliability, and safety. Our goal was to provide a single tool for prioritizing warnings and predicting potential issues.
ROLE
UX Interaction Designer
YEARS
2016 - 2019
Customer Interviews,
User-flow Mapping,
Information Architecture,
Empathy Mapping,
Design Sprint,
Mockups,
Prototypes,
Executive Presentations,
Final Specs
TEAM
Product Manager,
Subject Matter Expert,
Lead Architect,
Frontend Dev,
Backend Dev,
QA
DELIVERABLES
The Problem
Plant Operations Advisor (POA) was developed as a software tool for engineers to improve offshore oil platforms' efficiency, reliability, and safety. The objective was to create a product that could help engineers and BP overcome challenges like costly shutdowns that could amount to $10 million per day, difficulties in standardizing safety processes, and the need for a single tool to prioritize warnings and predict potential issues.
I oversaw UX design for events (Excursions and Anomalies) and case documentation (Cases) in the POA product. This involved user research, wireframes, and prototypes. I made decisions based on user needs, purpose, and testing. We launched and learned from the process.
These actions led to the successful launch of the POA pilot program, which became the first PREDIX application for external customers and sealed a $30M contract for GE.
As a result, the offshore reliability was increased from 92% to 96%, and the product was scaled to 87 engineering platforms.
What mattered to engineers?
To best understand the challenges our team faced in solving for our customers, I dove into learning as much as I could through stakeholder interviews, reviewing previous research reports, and spending time with subject matter experts.
User Discovery: team collaboration of persona empathy mapping and user journeys
Going further I worked with our team to synthesize our understandings. What were user's hopes, pains, and barriers to doing their jobs? What was important to gain trust and report with them?
Engineer pain points:
Data Silos
Lack of Real-Time Updates
Usability Issues
Inconsistent Data Interpretation
To many ways to pull data and know if an event is true: tools, email, ppt, etc.
Standardization
Lack of collaboration
Activities & Outputs
Customer Interviews
Stakeholder Interviews
User Journey & Empathy Mapping
Information Architecture
Learn, ideate, test, and improve
Launching our first pilot also required quick planning and consensus. The other design lead and I organized a framework that emphasized iterating and getting feedback from stakeholders and team members as quickly as possible in an agile environment.
We found that holding daily white-boarding sessions was a successful exercise in addressing business, technical, and user opportunities, risks, and constraints. By collaborating with a subject matter expert, product manager, and architect, we were able to make quick progress and reach a consensus on our solutions.
Iterative collaboration process: white-board ideas > wireframe > review > edit
I was able to take the sketches from our whiteboard sessions and quickly bring them into Sketch or Illustrator to create wireframes or modify existing designs. By utilizing a pre-existing design system, this process has helped me work more efficiently. I also made sure to tailor features to meet our specific use cases.
As part of our weekly process, we would test our workflows with users to ensure they meet their job requirements and for usability. At the end of the week, we would update our designs based on the feedback we received from our users and share them with our stakeholders.
Activities & Outputs
User Testing
Wireframing
Prototyping
UX Strategy
Interaction design: helping users go from A to B
To make the POA tool effective, data visualization played a crucial role. It was important to convey the information in a way that would guide users through each step of the workflow. To achieve this, I consulted with domain experts and users to ensure that the right amount of information was presented at each stage. It was essential to reveal the necessary data points without overwhelming users with excessive information. Additionally, allowing engineers to interact with and add data points increased their confidence in the software and data.
Data visualization interaction design exploration with Anomalies, level 1 data and level 2 data that allows engineers to interact with time series charts
To gain insight into how engineers interact with data, I created prototype data visualization interactions. These were designed based on how engineers use other tools or how they naturally interact with charts. One challenge was to avoid overcrowding the screen or making the process too complicated.
Designing case components and interactions
When developing the Case feature, it was crucial to include a variety of documentation and interactive components. Creating clear and informative messages and responses was a top priority. Additionally, I took great care to consider various use cases to ensure that the feature was easy to use and effective for engineers collaborating on diagnoses and solutions.
Excursion wireflow documentation
Managing the intricacies of POA was not an easy task. To overcome this, I developed intricate prototypes with multiple layers of user flows and annotations to cover various use cases and provide greater context to our PMs and developers.
Activities & Outputs
Wireframing
Prototyping
Final Specs
Stakeholder Presentations
Diagnose and document all in one
The POA pilot launched with both the Excursions and Cases features. Soon after we received some great feedback on how to improve both workflows. Overall, these features had over a 60% user satisfaction rating. With additional improvements like grouping events and usability improvements for Cases, that number increased to over 70%.
Diagnose an Excursion
Engineers can view all of the Excursions in their area. They can quickly scan and understand the priority of the event, when it happened, and if it’s still ongoing. Digging deeper they know what equipment the problem is happening and other important instrumental KPIs related to its performance. The data visualization provides a visual of the equipment’s performance and when the Excursion started. The blue box defines the event time, while the dotted lines let an engineer know the severity of the event. To provide greater context, a table provides a list of previous Excursions on the equipment.
Excursion design overview
Excursion inbox design
Excursion time series design
Grouping Excursions
For events that trigger additional Excursions downstream, Grouping allows engineers to group them together for a holistic approach to diagnosing a problem and creating a case. Engineers can easily add/remove other Excursions to a parent event.
Excursion grouping design: group and parent views
Case creation, collaboration, and evidence
Cases help engineers document and diagnose Excursion events by following a report workflow, commenting, assigning actions, and attaching evidence. Related cases also improves the knowledge sharing process between teams, while having the ability to upload files and screenshots from other monitoring systems provides a source of truth for teams.
Case design left: inbox, overview, report
Case design right: assigned Excursion, related events, and evidence
Case design evolution: v1 to v2 with improved jump navigation and inbox search capabilities
We reached 96% efficiency
We developed POA and its versions for three years. Excursion design was successful with an 80% satisfaction rate, while Cases had 70%. We launched on all 87 global platforms with a 96% efficiency rate. Decoding Anomalies was challenging, but the project was a great experience. We enjoyed fast decision-making and user testing, improving designs for better audience understanding.
Improved situational awareness from 92% to 96%
Decreased Excursion events with Grouping capabilities
Increased user satisfaction above 70% for Excursions and Cases
“We expect the deployment of this technology not only to deliver improvements in safety, reliability and performance of our assets, but also to help raise the bar for the entire oil and gas industry.”
Next Up
Solarity
Achieving customer acquisition goals in designing a movement!