GE Healthcare Life Sciences (now called Cytiva) hired me in 2019 to design the user interface for the software of their newest bioreactor. The interface of their other bioreactors looked dated and was cluttered with redundant and unnecessary information. Users were also experiencing alarm fatigue, which can result in missed alarms and alerts and could cost a company millions of dollars. Our bioreactor customers included biopharmaceutical companies like Merck, Pfizer, Amgen, Daiichi Sankyo and Genentech.
Scroll for a description of the UX process, user flows, wireframes, mockups and a prototype.
This project required me to learn HMI (Human Machine Interface) which is the model used for automation processes in plants and labs across many industries. HMI is different from UX and its guidelines often conflict with UX standards. An example of this would be the HMI guideline that says backgrounds should be gray with darker gray text over it to create a very low contrast display. The UX Designer in me would never recommend that so concessions and compromises between UX and HMI needed to be carefully considered. Sometimes UX won, sometimes HMI won. I also needed to familiarize myself with the possibilities and constraints of Rockwell FactoryTalk, which was the platform that the developers were using to build it.
Wireframes & User Flows
There was a lot of information, content, and science to figure out. I met with scientists and subject matter experts within the company and leveraged some of their relationships with outside organizations to work out Information Architecture and user flows. I determined what features and KPIs need to be on the screen at all times. We whiteboarded a lot, which led to loosely sketched wireframes like the ones you see below. These were reviewed, tested and refined until they were ready to be polished and included in prototypes.
Once we had a reasonable user flow and wireframes, a rough prototype was built in inVision. The idea was to get something in front of people quickly and get feedback. The data doesn't lie, so the sooner you can get some tests with real users, the better. While I always hope early feedback will be positive, I welcome criticism and surprises. Testing shows us where the design can be improved.