Seahorse Analytics Plate Assay Builder
This is a "Case Study" that was created for marketing our department to the company at large. It provides a good example of the benefits of UI/UX design in general and how it can be integrated into an Agile development process. It also showcases the design process that I led in the creation of this patent-pending UI element.
The web app is seahorseanalytics.agilent.com and is free to sign up and use.
Executive Summary
The first part of the project was to examine the known usability issues of the existing assay design workflow that was part of the Wave desktop software. The second part was to reimagine an entirely new, yet familiar software interface for the Seahorse Analytics web platform that would help to eliminate those issues. This was a part of an overall move from a PC-based desktop software platform (Wave) to a more agnostic web-based interface.
Project Timeline
Background
We knew from feedback from the field that most users don’t use the included templates to design their assays in Wave. Because the templates required users to manually add repetitive information, they would just skip the templates and enter the bare minimum to get the experiment to run. Ultimately, this behavior causes a barrier to gaining insights when analyzing the assay results. Clearly there had to be a better way.
Users find their own path if the solution doesn't work for them.
This was the assay creation process we were looking to improve. Even though templates came preloaded, users still didn't use them.
Case Evaluation
We gathered feedback from FAS’s (Field Application Scientists), support and sales. We asked our internal bio apps teams how they think about how assays are designed based on biology. We ideated and came up with several ways to turn the feedback into a design solution.
The goals:
Eliminate the need for the user to enter repetitive data and thereby eliminating errors associated with it.
Greatly shorten the time required to design an assay/experiment.
Give users a clear visualization of the location of individual parameters on the plate.
A 96 well plate used across cell analysis.
It is important to note that these exercises were not extensive and did not delay the project in any way. In fact, stakeholders acknowledged that the process actually ended up saving time and that they wouldn't want to develop future projects without them.
Proposed Solutions / Present Results
We took existing prototypes from Wave (the previous desktop software) and reused the parts to quickly build a rough design sketches. This allowed us to iterate quickly, developing multiple design approaches to gain early feedback from the stakeholders.
Once a design direction was chosen, we developed high-resolution interactive prototypes. These prototypes allowed us to test out the intended novel functionality early, prior to involving software dev. We could gather feedback, identify potential usability issues, make adjustments to the design quickly and retest. We even used our final prototype in usability testing with customers…all prior to any coding.
Usability Test
Remote usability tests amid the pandemic
Planning for our usability tests began prior to the pandemic. However, because our users were spread around the globe, we had already planned for remote testing via MS Teams. It was lucky that we did, as our tests ended up happening in June of 2020. We needed to navigate the sudden change of availability of participants operating in biology labs. Users performed the tests on their choice of computer at work or home using a web browser and a shared link. All tests were recorded with screen sharing. The yet-to-be-coded UI performed better than we expected and helped reinforce that the new UX design was valid.
Afterwards, we brought the validated designs to software dev for coding. We continued to work with the team throughout the software development process to ensure alignment to the design criteria.
Planning for the unplanned
There are many instances and interactions that you cannot plan for in a non-functional interactive prototype. “What if the user does this?” is a question that can often create holes in a beautifully planned design. However, working hand-in-hand with the software development team allowed us to come up with innovative solutions that solved problems while still maintaining the intended functionality.
An example of this is the Auto Cleanup feature for the groups. If users navigated the software in an unexpected manner, it could cause the unintentional creation of duplicate groups. This feature scans for these accidental groups and removes them before they can cause any confusion.
Figure 1 Auto-cleanup feature UI
The Final Solution
A (very) short demo showing just how quickly a user can design an assay using the Plate Assay Builder…in this example, under a minute. Although most users would want to add more details and custom naming, it's more of a demo to show how interactions with the Plate Map dynamically build complex groups.
Conclusion
The final interface, which we call the Plate Assay Builder, fulfills all the user needs identified early in the development process.
Goals
Solutions
Eliminate the need for the user to enter repetitive data and thereby eliminating errors associated with it.
Parameter details are only entered once. When applied to the plate, the groups are automatically created. Changes to parameters are automatically updated in all places
Greatly shorten the time required to design an assay/experiment.
The solution dynamically creates assay groups based on the different parameter combinations in real-time, eliminating several steps from the original method. Although not officially tested for speed improvements, observations of users has shown a dramatic reduction in assay creation time.
Give users a clear visualization of the location of individual parameters on the plate.
Directly applying the parameters to the plate, rather than to groups, solves this problem. Additionally, it more closely aligns to how users conceptualize their assay.
Overall, the project went quite well and all stakeholders were very happy with both the process and the results. The team acknowledged the advantages in cost savings and improved product design that accompanied this collaborative effort by following the Human Centered Design process. Most importantly, customers have been adopting this new assay creation method and greatly appreciate the improved experience.
Invention Disclosure and Patents
Our team felt it was such a unique approach that we filed for an Agilent Invention. The User Interface is now patent pending. Fast forward now to 3 years later…having identified numerous potential improvements and features, the next generation will be even more refined. We plan to file for a patent revision to cover these improvements.
Quotes
“When I was presenting this to a new beta tester, before I could finish, they said out loud, ‘I already know how this works’…the customer finished creating their template using this new software application that they had never been trained on before, using new features without any step-by-step guide or anything."
“I think that our software development would be slower if we didn't use high fidelity prototypes. When we don't even have to code it…it gives a sense of what the software experience would be before we actually spent time and energy and money to create those features.”
–Ben Paradis, Product Manager
Interview with Product Manager