Back to all projects

Project type

SaaS, MVP

Timeline

2 weeks

My Role

UX/UI design

ROAS AI

ROAS AI

A SaaS platform built to help Shopify e-commerce brands enhance their Meta advertising campaigns through AI-driven signal optimization instead of relying on manual bidding.

Goal

Goal

Create a neat and clean MVP solution for a wide audience of Shopify merchants, who are not marketing specialists and would like to improve their Meta ads results witouth diving into targeting details.

Challenge

Shopify merchants often struggle to scale Meta ads efficiently

The process of identifying high-value audiences, predicting LTV, and continuously optimizing campaign targeting is time-consuming, technical, and often based on guesswork

Solution

Platform that could be used without needing to manually manage campaigns

Complex AI behavior feeling understandable and trustworthy

Design proccess

Design proccess

1

Discovery & Research

  • Problem definition

  • Stakeholder’s research

  • Competitors analysis

2

Defining

  • Information architecture

  • Userflows

  • Prototyping with Claude AI

  • Low fidelity Sketching & Wireframes

3

UI kit

Built with shadcn/ui for speed and design consistency, with minor custom styling for brand alignment

4

High fidelity design

5

User Testing & Feedback

Usability testing*

*accomplished for learning purposes due to lack of time and resources during the development

6

Development handoff

Campaigns

Campaigns

Create, manage, and activate custom optimizations that power smarter Meta ad targeting without manual guesswork.

Campaign creation features a streamlined view to help users achieve first results fast, with advanced settings available for deeper research and optimization.

Retargeting

Retargeting

A specific type of advertising strategy where ads are shown to people who have previously interacted with the business but haven’t converted yet.

I’ve decided to separate retargeting from regular campaigns to simplify the campaign creation view and better reflect the different mental models users have when working on them.

It’s the difference between: “Let me set up and refine something new” vs “Let me check how my funnel recovery is working”

A specific type of advertising strategy where ads are shown to people who have previously interacted with the business but haven’t converted yet.

I’ve decided to separate retargeting from regular campaigns to simplify the campaign creation view and better reflect the different mental models users have when working on them. It’s the difference between: “Let me set up and refine something new” vs “Let me check how my funnel recovery is working”

  • Performance summary cards (ROAS, AI training status, active signals, ROI)

  • Revenue vs Ad spend timeline chart

  • Business impact metrics

Learnings and Future Improvements

Learnings and Future Improvements

Working on this project gave me the opportunity to dive deep into the nuances of Meta ads marketing and performance.

With more flexible timeline and budget, I would have added the following to the process:

User interviews during the research phase to better understand user struggles and mental models

Usability testing of competitor interfaces (if needed based on interview insights) to identify blind spots in existing solutions

Usability tests with the target audience, followed by design refinements based on their feedback

As the product grows and we gather user feedback, it will continue to evolve. I've already identified several improvements for future consideration:

Simple, user-friendly onboarding experience

Real-time AI health and learning monitoring indicators

Pre-built AI recommendation templates for different goals and use cases

Thanks for

watching!

Thanks for

watching!