Case Study: AI-Generated Product Photo App – Beta Launch & Market Validation

Role: Senior Product Manager
Duration: August 2022 – November 2024

‍

AI App improves marketing images based on user prompt "Add violet background and herbs around the cosmetic tube."

Project Overview

The objective of this project was to develop and launch a beta version of an AI-powered product photo generation app for eCommerce sellers. The app allowed users to create professional product images using either pre-designed templates or custom AI-generated prompts.

The primary goal was to validate market demand, user adoption, and conversion potential through structured product testing and targeted advertising.

Key deliverables included:

  • A fully functional beta version with core AI image generation capabilities.
  • A Google Ads-driven go-to-market strategy to measure direct user acquisition and conversion rates.
  • In-app purchase flows requiring users to connect a credit card for a 10-day free trial before converting to a paid subscription.

Challenges & Initial Observations

Low Prompt Usability:

  • The first version of the app only allowed custom AI-generated prompts, but users struggled to write effective prompts.
  • Poorly written inputs resulted in low-quality image outputs, reducing user satisfaction and retention.

Engagement & Retention Issues:

  • Only 1% of trial users converted to the paid plan, signaling a need for better onboarding and guidance.
  • High subscription cancellation rates suggested that users were unsure of the app’s long-term value.

Paid Acquisition Efficiency:

  • Google Ads successfully drove signups, but the 4% conversion rate from ad clicks to signups was not translating into paid retention.
User Signup Onboarding
User Signup Onboarding - 7 days free trial
User Signup Onboarding - Stripe Credit Card Connect

🚀 Agile Execution & Jira Workflow

To ensure efficient execution, I led the project using an Agile methodology, focusing on structured sprints and iterative improvements.Agile Process & Sprint Structure

  • Sprint Planning: Defined Epics (e.g., "Guided Prompt Optimization", "Template-Based UX Enhancements") and created actionable Jira tickets.
  • User Story Development: Prioritized key user journeys, ensuring that features addressed core pain points.
  • Backlog Grooming: Regularly refined backlog priorities based on user feedback and data insights.
  • Development & QA Sprints: Ensured bi-weekly releases, allowing continuous iteration.

Solution: Improving User Experience Through Structured GuidanceAdding Template-Based AI Generation:

  • Introduced pre-built templates for product categories (watches, perfumes, electronics) to eliminate guesswork for users.
  • Outcome: Enabled faster and more accurate image generation, significantly improving user success rates.

Implementing Guided Prompts for Custom Input:

  • Added a real-time guided prompt assistant, providing users with suggestions while typing.
  • Outcome: 80% of users successfully generated usable images, compared to almost none before.

Refining UI & User Flow:

  • Improved navigation and feature visibility, ensuring that advanced settings were clearly discoverable from the start.
  • Outcome: Reduced user frustration and confusion, increasing engagement time in the app.

‍

Initial Product List Page
Initial Product Image Generation Page
Upload Image to be generated

‍

User Testing & Data-Driven Iteration

Research & Testing Approach:

  • User Interviews: Conducted structured feedback sessions with early users to identify UX pain points.
  • Behavioral Analytics (PostHog): Set up event tracking for each user journey, monitoring drop-off points and feature adoption rates.
  • A/B Testing: Compared before and after engagement rates with templates vs. custom prompts.

Live Testing & Iteration:

  • Outcome: Post-launch analysis showed that the new guidance systems significantly improved initial user satisfaction, though retention beyond the free trial remained a challenge.

Go-To-Market Strategy & Conversion Metrics

Google Ads for Market Validation:

  • Implemented a strictly targeted keyword strategy to attract high-intent users.
  • Used detailed funnel tracking to measure conversion from ad click → signup → paid user.

Conversion Flow:

  • Users were required to connect a credit card upon signup for a 10-day free trial.
  • 4% of visitors signed up, but only 1.5% converted to a paid subscription.
  • Despite a 40% revenue recovery on ad spend, further optimizations were needed to improve long-term retention.

Metrics Tracking in PostHog:

  • Measured every step of the user flow, tracking which ad keywords led to the highest conversion rates.
  • Identified which features users engaged with the most, guiding future product improvements.

📌 Key Results

  • 4% conversion from paid ads to signups.
  • 80% success rate in generating high-quality AI product images after introducing guided prompts.
  • 1.5% conversion rate to paid plans, proving market willingness to pay, but highlighting a need for retention improvements.
  • 40% marketing budget recovered in initial revenue, validating the business concept but requiring further refinements.
Posthog metrics app conversions based on user Google Ads keywords
Posthog Google Ads onboarding conversion rates

‍

Key Learnings & Recommendations

Improve User Guidance Further:

  • While templates and guided prompts significantly improved user success rates, further onboarding enhancements were needed to ensure users understood the full capabilities of the app.

AI Model Optimization:

  • Enhancing AI prompt interpretation and image quality (potentially transitioning to Stable Diffusion) would increase user trust and subscription retention.

Refine Monetization Strategy:

  • Consider alternative pricing models such as lower-cost monthly plans or usage-based pricing to improve retention and reduce cancellations.

My Contributions

  • Led end-to-end product development, from concept validation to beta launch.
  • Defined and executed Agile sprints, managing tasks in Jira across design, engineering, and marketing teams.
  • Implemented user guidance improvements, introducing templates and real-time prompt assistance.
  • Developed a data-driven iteration process, leveraging PostHog analytics and user interviews to optimize the product experience.
  • Designed and measured a go-to-market strategy, ensuring data-backed decision-making for ad spend efficiency.
  • Successfully validated a startup concept, proving real user willingness to pay for AI-powered eCommerce product images.

Final Outcome & Next Steps

This project successfully validated market interest in AI-powered product image generation. Despite strong initial engagement, long-term retention and monetization required further refinements.Next Steps

  • Deepen AI functionality: Enhance model capabilities for higher-quality, more accurate image generation.
  • Expand onboarding support: Introduce interactive onboarding to ensure users fully understand the product.
  • Optimize pricing model: Experiment with alternative pricing structures to increase paid retention.

By leveraging user data, iterative testing, and Agile execution, this project demonstrated strong product-market potential and provided valuable insights for scaling the business.