Redesigning a visual analysis platform
At Pytri, I led the end-to-end design of a platform that helps microbiologists measure bacterial colony growth under varying environmental conditions. My process included conducting user research, identifying three core growth principles, ideating solutions, wireframing, prototyping, running usability tests, and handing off final designs to developers.
Impact
100K+ Funding Granted for uncovered value proposition
65% Increase in User Signups
Solution Preview
Pytri is an AI-powered image analysis platform that lets microbiologists upload lab images and instantly receive assay counts. By automating detection, it eliminates manual counting and saves hours per sample while ensuring fast, consistent data.
This project highlights how our team redesigned the platform to surface new value propositions and increase retention. Heres a short preview of some the solutions we implemented.
The Challenge
Hired to solve a growth problem: retention
Pytri had strong adoption among microbiologists, but retention dropped off quickly. Most users exported their results after a single experiment and never returned. The platform wasn’t embedded in their ongoing workflow, limiting long-term engagement and growth.
Defining the problem space
Conducting research and understanding user needs
I interviewed 15 users and mapped insights into an ideal persona. I created a research question for her and simulated her workflow from experiment setup to results. This helped identify points where Pytri could grow by integrating seamlessly into her workflow routines.
Meet Dr. Elena Rao
Role: Post Doctoral researcher in Microbial Resistance
Institution: University hospital infectious disease lab
Experience: Comfortable with digital lab notebooks, basic scripting (Python/R), statistical software (GraphPad, SPSS)
Research Question:
Do bacteria grow differently when exposed to different types of stress, like UV light, heat, or pH changes?
From here I walked through the process of our user persona conducting this experiment from start to finish identifying where Pytri can integrate into her workflow along the way.
Key Insights & Ideation
Users return when their data lives inside the platform
Through user interviews and workflow analysis, I uncovered a key principle for retention:
Users return when their data is easy to access, revisit, and build on.
Researchers stated that they valued revisiting and building on past work, but Pytri wasn’t enabling that behavior. If Pytri remained a one-time analysis tool, retention would always plateau.
My team and I brainstormed solutions around three levels:
Lightweight: Nudges, reminders, and habit loops.
Mid-level: Saved workflows, dashboards, experiment history.
Bold bets: Interactive datasets and auto-generated reports.
Idea Validation
Initial explorations with interactive datasets
I hypothesized that interactive datasets and reports would anchor users. Since this idea required a heavy development effort, I created mockups and tested them to validate the ideas first.
Result: Users liked the concepts but adoption issues persisted. Many didn’t realize their data was already stored in Pytri and assumed it disappeared after export.
This revealed a key flaw: retention issues weren’t only about new features but about basic awareness of existing value.
Finalizing solutions
Pivoting from bold bets to practical fixes
Armed with new insights, I shifted focus toward mid-level solutions that clarified existing value before adding more complexity.
These included:
Improved information architecture with collapsible navigation bars and breadcrumbs..
Guided upload user flows for naming, editing, and saving projects.
Uncovering collaboration features as a primary action for easy sharing.
These solutions aligned with a core growth principle: before adding new value propositions, ensure users recognize the existing ones.
→
Final Designs
From Concepts to Execution
After aligning on solutions, I created wireframes and high-fidelity designs, running weekly critiques with the team to gather feedback and refine ideas
Familiar Navigation Patterns
I added a collapsible navigation bar, breadcrumbs, and project naming during uploads to improve visibility and recognition. This made it easy for users to locate and return to their data, which reduced friction and encouraged repeat use by clarifying the value that already existed in the platform.
Uncovered collaboration features
I surfaced collaboration as a primary action on project pages, allowing users to share data with colleagues instantly by adding emails and setting permissions. This made sharing effortless and positioned Pytri as a central hub for stored data, creating social accountability and stronger reasons for users to return.
Simplified upload flow
I designed a guided upload flow that required users to name, edit, and publish projects before saving. By personalizing their data upfront, users felt more ownership and attachment to their projects, which made stored results easier to recognize and encouraged them to return and reuse the platform.
Reflections
Lessons in growth design
Hypotheses need testing: My first instinct was wrong, but testing saved my team from investing heavily in features users weren’t ready for.
Clarity before complexity: Sometimes retention isn’t about bold new features, it’s about making sure users understand the value that already exists.
Growth is iterative: Retention challenges rarely have a silver-bullet solution. It’s about layering quick wins and learning toward bigger bets.
