Pitney Bowes, a print mailing solutions leader, has high SLA penalties for every second of machine downtime. These machines have hundreds of moving parts and whenever a stoppage alarm is triggered, the service technician is tasked with identifying, diagnosing and resolving the issue with the goal to minimize downtime. The Clarity™ Advisor solution, built on GE Digital's Predix platform, is essentially a maintenance and diagnostics tool catered to the print mailing industry.
Role and Responsibilities
I inherited this project from a senior designer when the application was in version 1.3 and I lead all design initiatives for three major release cycles. My main responsibilities have included:
User Research: Collaborating with primary and secondary users at client's site to understand pain points.
Interaction Design: High level design strategy, UI framework and wireframing, Prototyping in InVision as well as JSFiddle and stakeholder presentations.
Visual Design: Designing production-ready visuals for the application.
Discovery and Exploratory Research
- I quickly got up to speed, meeting with stakeholders to validate the previous work and clean up technical debt while ensuring project roadmap deadlines were being met. Usability enhancements to key features, e.g. selection of favorites, were made during the initial release.
- The concepts created by previous designers were done under tighter deadlines and lacked documentation; so I mapped the current state workflow to drive team alignment.
- My product manager and I visited the Pitney Bowes headquarters in Danbury, CT for a two day discovery workshop with various stakeholders.
- Business Stakeholders (including SMEs) – to understand key requirements and roadblocks.
- Service Technicians – Brainstorming and asking questions about how the app will be used, deep dive into features and wireframes.
- Engineering – to understand security and feasibility of design ideas.
- I also observed users in their real working environment and studied how they would interact with our product in order to achieve their goal of maximizing machine uptime.
Concept and Generative Research
A key feature of the new release was the introduction of a KPI around volume of mail pieces and the goal was to eliminate low volume jobs from the data visualizations. I did research on what visualizations existed from a design system perspective, but not much of it was in landscape. So I wireframed layout options and carried out internal and external design reviews for the three different options.
- Exploration #1 - Introduction of a new component (Analysis Slider) which eliminated alarms of low production.
- Exploration #2 - Introduction of a new feature (Heatmap) within the existing efficiency table which highlighted high production alarms. There were two subversions within this exploration, each of which used different metrics as primary indicators.
- Exploration #3 - Introduction of a simplified data point (Loss/Gain) which replaced multiple data points (of efficiency and/or volume); The heatmap from exploration #2 was leveraged for this concept.
My success metric for these explorations was the findability of the most important alarm for the service technician to fix. I carried out internal design reviews as well as collected feedback from the client and v2 within Exploration #2 emerged as the clear winner.
Refinement and Evaluative Research
I fleshed the chosen concept by creating detailed visual specifications and defining key interactions. As the developers were bugfixing and cleaning up some of the scalability issues for the web app, I took the opportunity to share my designs with the broader design community. The Predix Design team was holding sessions to gather requirements from other business units as to how they use the system. I used the sessions as a way to gather ideas and also contributed my components to the Predix Design System.
Once the new features were deployed in the testing environment, I carried out knowledge sharing sessions around best practices for conducting usability evaluations with their clients. Some of the key pieces of feedback were around:
- Improving performance issues.
- Standardizing naming conventions for different components in the feature overlay/coach marks component.
- The alarm chart, with its enhanced timeframe views, being a great delighter.
Reflection and Next Steps
GE Digital is an engineering-driven company where most of the software engineers I've worked with are reticent and prefer working in their siloed space. In order to ensure that the integrity of designs was maintained in implementation, I made an extra effort to speak their language. I worked very closely with all front-end engineers ensuring the specifications were being delivered in their tool of preference - while some components were well received as InVision prototypes, other data visualization components were preferred as JSFiddle code. Learning to work with a distributed team added an additional challenging facet to this product given that the client was on a different time zone. Communication, documentation, and patience became key elements for success.
Advisor will continue to develop as a real-time monitoring and diagnostics tool for the Emerging Verticals team at GE Digital. Some of the next steps include design of email notifications as well as the enhancement of the site view page to provide a deeper dive into machine-specific metrics.