Back

Back

[Case 01]

Enhancing Cell Assay Workflows

BioTech

Enhancing Cell Assay Workflows

[Project Overview]

Optimizing Cell Assay Workflows from expanding 96-Well Plate Map Setup design to 384-Well Plate Map design on the Promega ProNect™ Data Platform.

This project is still ongoing and will be updated soon with clickable demo results, improved concepts, and post demo survey results.

[User Research]

How might we extend the ProNect Data Platform to support both 96-well and 384-well plate formats so that scientists can seamlessly analyze experimental data across different plate types? 

Promega is a global biotechnology company that provides innovative solutions for life sciences. CellTiter-Glo is one of its flagship assays used to measure cell viability, and ProNect is a digital platform designed to streamline assay setup and data management workflows.

As part of a cross-functional initiative at Promega, I led UX design research to reimagine digital tools for 384-well plate mapping. Our goal was to improve workflows for scientists working in high-throughput areas like genomics, drug discovery, and cell-based assays by making plate setup more intuitive, accurate, and scalable. I built off the existing 96-well plate design to introduce dual-compatibility for more complex 384-well plate maps.

Industry

BioTech

My Role

UI/UX Design Intern

Platforms

Desktop

Timeline

July 2025 - Present

ProNect Data Platform

[Persona]

Ellie Roberts

Research Scientist (High-Throughput Screening)

Research Scientist (High-Throughput Screening)

Research Scientist (High-Throughput Screening)

I thrive on precision and efficiency

Age: 29

Location: New York City

Tech Proficiency: Moderate

Gender: Female

[Goal]

Throughput & Scale: Handle large sample volumes without sacrificing accuracy.

Reproducibility: Ensure results are consistent across runs and operators.

Designs Experiments: Sets up hundreds of conditions in parallel to test hypotheses quickly.

[Frustrations]

Manual setup bottlenecks in large-scale experiments.

Setup complexity — higher density means more chances for human or robotic error.

Poor scalability: Tools aren’t optimized for handling large plate volumes or repeated layouts.

[User Research]

To build a foundational understanding of user workflows and preferences, we adopted a two-phase approach that began with qualitative discovery and transitioned into broader quantitative validation.

Phase 1: Internal Scientist Interview:

We conducted a one-on-one interview with a Promega scientist deeply familiar with high-throughput workflows and 384-well plate usage. This session helped us map out the typical end-to-end process: from designing plate layouts to integrating with instruments and analyzing outputs.

Phase 2: Survey Design & Client Outreach.

Building on our interview findings, I drafted a set of 15 potential survey questions to probe deeper into user roles, throughput levels, software habits, and pain points. After refining the content with our team, we distilled the survey down to 5 essential questions to reduce friction and increase response rate.

To build a foundational understanding of user workflows and preferences, we adopted a two-phase approach that began with qualitative discovery and transitioned into broader quantitative validation.

Phase 1: Internal Scientist Interview:

We conducted a one-on-one interview with a Promega scientist deeply familiar with high-throughput workflows and 384-well plate usage. This session helped us map out the typical end-to-end process: from designing plate layouts to integrating with instruments and analyzing outputs.

Phase 2: Survey Design & Client Outreach.

Building on our interview findings, I drafted a set of 15 potential survey questions to probe deeper into user roles, throughput levels, software habits, and pain points. After refining the content with our team, we distilled the survey down to 5 essential questions to reduce friction and increase response rate.

To build a foundational understanding of user workflows and preferences, we adopted a two-phase approach that began with qualitative discovery and transitioned into broader quantitative validation.

Phase 1: Internal Scientist Interview:

We conducted a one-on-one interview with a Promega scientist deeply familiar with high-throughput workflows and 384-well plate usage. This session helped us map out the typical end-to-end process: from designing plate layouts to integrating with instruments and analyzing outputs.

Phase 2: Survey Design & Client Outreach.

Building on our interview findings, I drafted a set of 15 potential survey questions to probe deeper into user roles, throughput levels, software habits, and pain points. After refining the content with our team, we distilled the survey down to 5 essential questions to reduce friction and increase response rate.

To build a foundational understanding of user workflows and preferences, we adopted a two-phase approach that began with qualitative discovery and transitioned into broader quantitative validation.

Phase 1: Internal Scientist Interview:

We conducted a one-on-one interview with a Promega scientist deeply familiar with high-throughput workflows and 384-well plate usage. This session helped us map out the typical end-to-end process: from designing plate layouts to integrating with instruments and analyzing outputs.

Phase 2: Survey Design & Client Outreach.

Building on our interview findings, I drafted a set of 15 potential survey questions to probe deeper into user roles, throughput levels, software habits, and pain points. After refining the content with our team, we distilled the survey down to 5 essential questions to reduce friction and increase response rate.

Foleon Survey we sent to cilents

[Insights]

We received initial feedback from scientists with varying levels of plate throughput (1–10 plates per week).

Tools used:

Excel and GraphPad Prism were the most common. What works: Excel is accessible and easy to edit.

Pain points:

Excel setup is extremely manual, especially for 384-well formats. Users must build 16x24 grids and label wells by hand. Engagement: Both respondents expressed interest in participating in future research.

We received initial feedback from scientists with varying levels of plate throughput (1–10 plates per week).

Tools used:

Excel and GraphPad Prism were the most common. What works: Excel is accessible and easy to edit.

Pain points:

Excel setup is extremely manual, especially for 384-well formats. Users must build 16x24 grids and label wells by hand. Engagement: Both respondents expressed interest in participating in future research.

We received initial feedback from scientists with varying levels of plate throughput (1–10 plates per week).

Tools used:

Excel and GraphPad Prism were the most common. What works: Excel is accessible and easy to edit.

Pain points:

Excel setup is extremely manual, especially for 384-well formats. Users must build 16x24 grids and label wells by hand. Engagement: Both respondents expressed interest in participating in future research.

We received initial feedback from scientists with varying levels of plate throughput (1–10 plates per week).

Tools used:

Excel and GraphPad Prism were the most common. What works: Excel is accessible and easy to edit.

Pain points:

Excel setup is extremely manual, especially for 384-well formats. Users must build 16x24 grids and label wells by hand. Engagement: Both respondents expressed interest in participating in future research.

[Design Solution #1]

Zoom Navigation for Dense GridsTools used:

One of the first pain points identified was the difficulty of navigating large 384-well plate layouts in Excel. Users often struggled to manually manage 16x24 well grids, making it hard to locate, edit, or review well-level data.

To address this, I introduced a zoom and quadrant-based navigation system to break the grid into manageable sections. This allowed users to: Zoom out to see the entire plate for a quick overview Zoom into individual quadrants (A–D) for focused editing, and navigate through quadrants using labeled tabs for clarity and speed. This model dramatically reduced cognitive load while maintaining flexibility for users working at different scales.

Zoom Navigation for Dense GridsTools used:

One of the first pain points identified was the difficulty of navigating large 384-well plate layouts in Excel. Users often struggled to manually manage 16x24 well grids, making it hard to locate, edit, or review well-level data.

To address this, I introduced a zoom and quadrant-based navigation system to break the grid into manageable sections. This allowed users to: Zoom out to see the entire plate for a quick overview Zoom into individual quadrants (A–D) for focused editing, and navigate through quadrants using labeled tabs for clarity and speed. This model dramatically reduced cognitive load while maintaining flexibility for users working at different scales.

Zoom Navigation for Dense GridsTools used:

One of the first pain points identified was the difficulty of navigating large 384-well plate layouts in Excel. Users often struggled to manually manage 16x24 well grids, making it hard to locate, edit, or review well-level data.

To address this, I introduced a zoom and quadrant-based navigation system to break the grid into manageable sections. This allowed users to: Zoom out to see the entire plate for a quick overview Zoom into individual quadrants (A–D) for focused editing, and navigate through quadrants using labeled tabs for clarity and speed. This model dramatically reduced cognitive load while maintaining flexibility for users working at different scales.

Zoom Navigation for Dense GridsTools used:

One of the first pain points identified was the difficulty of navigating large 384-well plate layouts in Excel. Users often struggled to manually manage 16x24 well grids, making it hard to locate, edit, or review well-level data.

To address this, I introduced a zoom and quadrant-based navigation system to break the grid into manageable sections. This allowed users to: Zoom out to see the entire plate for a quick overview Zoom into individual quadrants (A–D) for focused editing, and navigate through quadrants using labeled tabs for clarity and speed. This model dramatically reduced cognitive load while maintaining flexibility for users working at different scales.

[Design Solution #2]

Plate Type Flexibility

Dual compatibility for 96-well plates. Users could now toggle between 96- and 384-well modes using a simple selector, ensuring the tool could support both high-throughput and standard lab workflows.

Designed a toggle because there will only be 2 modes

Plate Type Flexibility

Dual compatibility for 96-well plates. Users could now toggle between 96- and 384-well modes using a simple selector, ensuring the tool could support both high-throughput and standard lab workflows.

Designed a toggle because there will only be 2 modes

Plate Type Flexibility

Dual compatibility for 96-well plates. Users could now toggle between 96- and 384-well modes using a simple selector, ensuring the tool could support both high-throughput and standard lab workflows.

Designed a toggle because there will only be 2 modes

Plate Type Flexibility

Dual compatibility for 96-well plates. Users could now toggle between 96- and 384-well modes using a simple selector, ensuring the tool could support both high-throughput and standard lab workflows.

Designed a toggle because there will only be 2 modes

[Design Solution #3]

Enhancing Grid Legibility

Added visual quadrant dividers (yellow lines) to segment the 384-well grid into four logical sections (A–D), helping users quickly orient themselves and reduce cognitive load when scanning or editing dense data layouts.Designed a toggle because there will only be 2 modes

Introduced a quadrant carousel navigation at the bottom of the interface, allowing users to quickly recognize what grid they are viewing and jump forward and backward between plate sections without losing track of their location.

Enhancing Grid Legibility

Added visual quadrant dividers (yellow lines) to segment the 384-well grid into four logical sections (A–D), helping users quickly orient themselves and reduce cognitive load when scanning or editing dense data layouts.Designed a toggle because there will only be 2 modes

Introduced a quadrant carousel navigation at the bottom of the interface, allowing users to quickly recognize what grid they are viewing and jump forward and backward between plate sections without losing track of their location.

Enhancing Grid Legibility

Added visual quadrant dividers (yellow lines) to segment the 384-well grid into four logical sections (A–D), helping users quickly orient themselves and reduce cognitive load when scanning or editing dense data layouts.Designed a toggle because there will only be 2 modes

Introduced a quadrant carousel navigation at the bottom of the interface, allowing users to quickly recognize what grid they are viewing and jump forward and backward between plate sections without losing track of their location.

Enhancing Grid Legibility

Added visual quadrant dividers (yellow lines) to segment the 384-well grid into four logical sections (A–D), helping users quickly orient themselves and reduce cognitive load when scanning or editing dense data layouts.Designed a toggle because there will only be 2 modes

Introduced a quadrant carousel navigation at the bottom of the interface, allowing users to quickly recognize what grid they are viewing and jump forward and backward between plate sections without losing track of their location.

[Design Solution #4]

Hover Preview in Fit Mode

To enhance usability in the Fit (full plate) view, I introduced a hover-to-preview feature.

When users hover over a cell, a tooltip appears showing the well’s full metadata (e.g., compound, concentration, cell type).

Hover Preview in Fit Mode

To enhance usability in the Fit (full plate) view, I introduced a hover-to-preview feature.

When users hover over a cell, a tooltip appears showing the well’s full metadata (e.g., compound, concentration, cell type).

Hover Preview in Fit Mode

To enhance usability in the Fit (full plate) view, I introduced a hover-to-preview feature.

When users hover over a cell, a tooltip appears showing the well’s full metadata (e.g., compound, concentration, cell type).

Hover Preview in Fit Mode

To enhance usability in the Fit (full plate) view, I introduced a hover-to-preview feature.

When users hover over a cell, a tooltip appears showing the well’s full metadata (e.g., compound, concentration, cell type).

[Design Solution #5]

Concept B

Added flexible zoom levels to better support different user workflows. In addition to the full plate and single-quadrant views, users could now: Zoom into two quadrants at once (A+B or C+D) for improved context when reviewing or editing adjacent regions. Zoom further into a single quadrant, displaying only six columns at a time, to support high-precision input where detail and accuracy are critical, such as dose-response curves or replicates.When users hover over a cell, a tooltip appears showing the well’s full metadata (e.g., compound, concentration, cell type).

This tiered zoom system helped users maintain both a sense of spatial orientation and fine-grained control, without overwhelming the interface.

Concept B

Added flexible zoom levels to better support different user workflows. In addition to the full plate and single-quadrant views, users could now: Zoom into two quadrants at once (A+B or C+D) for improved context when reviewing or editing adjacent regions. Zoom further into a single quadrant, displaying only six columns at a time, to support high-precision input where detail and accuracy are critical, such as dose-response curves or replicates.When users hover over a cell, a tooltip appears showing the well’s full metadata (e.g., compound, concentration, cell type).

This tiered zoom system helped users maintain both a sense of spatial orientation and fine-grained control, without overwhelming the interface.

Concept B

Added flexible zoom levels to better support different user workflows. In addition to the full plate and single-quadrant views, users could now: Zoom into two quadrants at once (A+B or C+D) for improved context when reviewing or editing adjacent regions. Zoom further into a single quadrant, displaying only six columns at a time, to support high-precision input where detail and accuracy are critical, such as dose-response curves or replicates.When users hover over a cell, a tooltip appears showing the well’s full metadata (e.g., compound, concentration, cell type).

This tiered zoom system helped users maintain both a sense of spatial orientation and fine-grained control, without overwhelming the interface.

Concept B

Added flexible zoom levels to better support different user workflows. In addition to the full plate and single-quadrant views, users could now: Zoom into two quadrants at once (A+B or C+D) for improved context when reviewing or editing adjacent regions. Zoom further into a single quadrant, displaying only six columns at a time, to support high-precision input where detail and accuracy are critical, such as dose-response curves or replicates.When users hover over a cell, a tooltip appears showing the well’s full metadata (e.g., compound, concentration, cell type).

This tiered zoom system helped users maintain both a sense of spatial orientation and fine-grained control, without overwhelming the interface.

After a considerable amount of testing and design iterations we decided to add a mini map in concept B that can be turned on/off so the user can mentally visualize where they are located in the plate map:

[In Progress Client Demo]

[Outcome]

Reduced set up times by 75% and streamlining experimental configuration for life sciences researchers.
15/15 clients were satisfied with demo and rated both concepts highly
Iterations are still being made and will publically be launched in 2026 on the ProNect™ Data Platform.

[Key Learnings]

Simplifying complex workflows through design

Working with high-throughput 384-well plate mapping taught me how essential it is to create tools that balance precision with ease of use. Many scientists rely on general-purpose software that does not fully meet their needs, so thoughtful design can significantly improve their workflows.

Progressive disclosure and usability

I learned the importance of progressive disclosure in interface design. Features like zooming into plate sections and navigating at different levels of detail helped reduce cognitive load, making the experience less overwhelming and more efficient for users.

Turning insights into practical solutions

By translating qualitative feedback from scientists into features like quadrant navigation and hover previews, I developed solutions grounded in real needs. This process highlighted the value of collaboration with cross-functional teams and showed me how small, incremental improvements can meaningfully support scientific work.

Select this text to see the highlight effect