Meta AI's Imagine Home Page

Evaluative research study to understand which concepts best communicate value propositions for Meta AI's Imagine on mobile.

OVERVIEW

Most users who visited Meta Imagine left the page without interacting. Currently, users are unclear about when and why they should use Imagine products for expressive media.

To address this, I tested new landing page concepts to understand which ones successfully communicated a desirable use case of Imagine to users.

IMPACT

The research findings informed Imagine Home's final designs, highlighted the importance of guidance in new AI tools, and identified an opportunity for co-creative authentic image editing experiences with friends.

I have omitted or edited confidential information in this case study. All information in this case study is my own and does not reflect the views of Meta.

TYPE

Evaluative and Strategic Research

ROLE

UX Researcher
Core Team: Product Manager, UX Designer, SWE, Data Science, UX Researcher

METHODS

User interviews: usability testing, concept testing

TIMEFRAME

4 weeks

Background

Context: What is Meta AI's Imagine? Meta AI's Imagine is a tool that enables users to generate AI images using text prompts, and is accessible across Meta's family of apps. In this study, we are testing new concepts for the Imagine landing page on Instagram Direct Message.

Problem: Data science analysis showed that over 80% of users who currently enter the Imagine home page did not type a single character into the prompt bar before exiting. We hypothesized that users may not fully understand the value proposition and potential use cases of Imagine, therefore causing them to exit.

I tested a total of 3 concepts in two prototype versions, delivering actionable insights and identifying strategic opportunities for Imagine to a team of product managers, UX designers, data scientists, and engineers.

My Role: I designed the research plan, conducted usability and concept tests, extracted key findings, identified product opportunities, and delivered actionable product feedback alongside strategic recommendations.

My Impact

Sensitive details and information around research insights and strategic recommendations have been edited or omitted.

After synthesizing the research findings, I wrote a research report and created a presentation deck to share with my stakeholders. This deck and presentation fostered additional discussions on next steps and new opportunities. This research had the following impact:

Prioritized launching the Spotlight and Formats concepts that emphasized guidance and visual examples with a clear action button.
Turned 'bad news' on Concept 2 (a concept that did not resonate with users) into a new strategic opportunity on utility-based features that is currently being explored.
Provided the team with additional evidence to push for a pivot to focus on "fun" user education and impactful guidance for navigating new AI tools in Meta's family of apps.
Identified a compelling use case for group editing with AI tools, which could create unique and new ways for users to stay connected to and interact with their family/friends in their group chats.

Learnings & Reflections

Challenge: Simulating the AI experience for testing. Early testing for our AI products is often done with clickable prototypes and hefty manual prep to generate images internally, then link them to the correct prototypes. As a result, I collaborated closely with UX designers and engineers in Figma to help create and organize the personalized prototypes.
Challenge: The team could quickly jump to inaccurate conclusions with pure concept feedback, so I made sure to ground the insights around foundational learnings on mental models for AI creation and expressive media.
Challenge: Time. With more time, I would have worked with engineers and design to prepare a live prototype for testing on our test devices. This would better simulate a mobile experience (as compared to clicking through a Figma prototype).
Learning: Turn 'bad news' into actionable recommendations. When users did not find the Capabilities concept useful or clear, I highlighted the use cases that users found valuable and the design gaps that affected their perceptions. As a result, the team did not have to scrap a concept that they had been developing for months alongside Meta's EmuEdit models. The concept is currently being re-designed into another experience.
Learnings & Next Steps: Emerging use cases. This study highlighted the importance of grounding new users in potential use cases as they explore new tools. In next steps, the team and I continued our explorations into use cases with a power user study. This research is under NDA, but the findings solidified certain use cases that emerged over others and provided the team with direction on where to drill down in design and product strategy.

Meta AI's Imagine Home Page

Background

Goals

Research Questions

Research Plan

Recruit

Methodology

Timeline

My Impact

Learnings & Reflections

Return to Work

I'm always open to fun projects. Let's chat!