Overview
My role
Product designer
Services
Full design cycle: Product Design, Research, Design System
Platform
Desktop web
At a glance
Challenge 1
Before
Marketers struggled with complex, technical setup flows that required understanding AI model parameters.
After
User research revealed that marketers needed guided, step-by-step flows with clear context at each stage.
Use cases
Changes after the 1st usability test round
Contextual AI suggestions based on user input
Instead of showing generic recommendations upfront, the system now waits until users input their campaign goal and then suggests relevant variant types and offer structures tailored to that specific objective.
Progressive disclosure for technical controls
Moved advanced configuration options (frequency caps, custom rules) behind an expandable "Advanced settings" section. Marketers see a clean interface by default, while ML engineers can access full control when needed.
User feedback
Setting up a campaign used to take me an hour of trial and error. Now I breeze through it in 15 minutes because the flow is so straightforward and nothing feels hidden or confusing.
Sarah, lifecycle marketer at an enterprise company
It feels like the interface was designed by someone who actually runs campaigns. Everything is where I expect it to be, and I don't have to think too hard about what to click next.
David, marketing lead at an enterprise SaaS company
Usability testing insights
Question tested during usability interviews:
Will customers understand the Evergreen vs. Calendared split in the UI and know where to find a Calendared experimenter?
Does the Calendared tab meet customers’ needs?
Findings:
Both customers found the 3-column layout to be clear. One customer suggested that adding the audience size to the tile would be helpful, which we will plan to include for the MVP.
Both customers found and understood the Evergreen vs. Calendared tabs quickly and without questions.
About the feature
Marketers configure promotional offers with details like discount amounts and durations, creating trackable variables for AI experiments. The platform learns which offer types perform best with specific audiences, continuously optimizing future campaign decisions.
Users & roles
Marketers at enterprise companies
External client persona.
Machine learning integration engineers
Internal technical user persona.
The products & offers UX flow.
User feedback
The offer setup is way clearer now. I know what to fill in and why.
Jessica, email marketing manager at an e-commerce brand
I used to skip this step because I didn't get it. Now it's obvious how it helps personalization.
Mark, lifecycle marketer at a fintech company
Challenge 2
Conducting research
The PSSUQ (Post-Study System Usability Questionnaire is a 15-item standardized questionnaire. It is widely used to measure users’ perceived satisfaction of a website, software, system or product at the end of a study. PSSUQ originated from an internal IBM project called SUMS (System Usability Metrics) in 1988.
9
Number of participants
30
Mins interview
18
Open-ended questions
Diving deeper
Token types
