Scaling Mob: From Recipe Discovery to AI-Driven Meal Planner

Scaling Mob: From Recipe Discovery to AI-Driven Meal Planner

Strategic product pivot that drove 143% subscriber growth and defined the roadmap for 2025/26. A retrospective of my whole year at Mob split in 3 pillars: Innovation, Habit Formation, and Personalisation.

143%

Premium subscriber growth

143%

Premium subscriber growth

135.6%

Retention lift in batch cooking

135.6%

Retention lift in batch cooking

Established the design system

From a component library, to a functioning system, accelerating feature velocity across Web and Mobile

Established the design system

From a component library, to a functioning system, accelerating feature velocity across Web and Mobile

Physical cookbook from user interview
Physical cookbook from user interview

Pillar A: Innovation

Automated Meal Planner: Validating the "AI Assisted Meal Planning" direction with lean product design

We utilized "Wizard of Oz" spreadsheet testing to simulate full automation. While this MVP drove a 17.9% lift in trial conversion, follow up interviews revealed an underlying problem: users hesitated to commit because they lacked agency over the AI's decisions.

Smart Suggestions: Pivoting to an "AI Assisted Meal Planning" Model

Acting on the insight that Agency > Automation, we pivoted from "doing it for them" to "helping them do it." By giving users control over AI suggestions, we achieved a 19.4% conversion lift (beating the fully automated model) and validated that retention relies on Human-AI collaboration, not full automation.

Pillar B: Habit Formation

Batch Cooking: Cook a recipe once, enjoy it through the week

Users adopting Batch Cooking showed a massive 135.6% increase in retention (Day 8/9) compared to the baseline. We learned that users perceive recipes which can be easily stored and adjusted for variety as easier to meal plan with.

Pillar C: Personalisation

Personalised Discovery: Cemented the foundation of 3-stage personalisation engine

  1. Cold start phase: Gather and use user preferences

  2. Behavioral phase: Tailor the discovery feed based on behavioural insights as we shape the taste profile.

  3. Taste profile phase: Predictive suggestions based on cooking history and interests.

Contextual collections and improved search filters drove a 12.7% uplift in trial conversion, validating this direction.

Takeaways

Agency outperforms Automation

For deeply personal tasks like food, users prefer assistance over full automation. I found that AI drives higher retention when positioned as a force multiplier (helping users decide) rather than a replacement (deciding for them). The most successful AI interfaces preserve user agency.

Agency outperforms Automation

For deeply personal tasks like food, users prefer assistance over full automation. I found that AI drives higher retention when positioned as a force multiplier (helping users decide) rather than a replacement (deciding for them). The most successful AI interfaces preserve user agency.

Agency outperforms Automation

For deeply personal tasks like food, users prefer assistance over full automation. I found that AI drives higher retention when positioned as a force multiplier (helping users decide) rather than a replacement (deciding for them). The most successful AI interfaces preserve user agency.

The "Why" matters, transparency is a UX Feature

Personalisation without context feels random or invasive. By exposing the 'why' behind suggestions, we bridged the gap between the algorithm and the human. In AI-powered and personalised products, transparency is the primary driver of trust.

The "Why" matters, transparency is a UX Feature

Personalisation without context feels random or invasive. By exposing the 'why' behind suggestions, we bridged the gap between the algorithm and the human. In AI-powered and personalised products, transparency is the primary driver of trust.

The "Why" matters, transparency is a UX Feature

Personalisation without context feels random or invasive. By exposing the 'why' behind suggestions, we bridged the gap between the algorithm and the human. In AI-powered and personalised products, transparency is the primary driver of trust.

Create a free website with Framer, the website builder loved by startups, designers and agencies.