Shayne Heathfield
Case studies
About
Shayne Heathfield
Case studies
About
Health & Wellness
Driving growth with AI at the #1 health and fitness app with 94.6% conversion in food logs.
How I made logging food as easy as speaking—using LLMs to reach 300K+ users in year one, averaging 9 logs per week.
How I made logging food as easy as speaking—using LLMs to reach 300K+ users in year one, averaging 9 logs per week.
Industry
Industry
Health & Wellness
Health & Wellness
Health & Wellness
Project date
Project date
2024
2024
Platforms
Platforms
iOS, Android
iOS, Android
Hypothesis
We believe that by allowing users to log with their voice, the experience will be so fast and convenient that we will increase average number of foods logged per user.
Product goal
Reduce friction so that users engage with MFP more often, finding more value, and ultimately convert to a paid subscription.
KPI
Average number of foods logged per user, per day during the first 7 days after registration.
300k+
unique users in first year
unique users in first year
9 times
average usage per week
average usage per week
95%
conversion from query to log
conversion from query to log
To quickly validate real-world queries and speech-to-text, we built a voice logging chat experience which allowed us to launch our internal and beta testing rapidly.
To quickly validate real-world queries and speech-to-text, we built a voice logging chat experience which allowed us to launch our internal and beta testing rapidly.
Design leadership
Design leadership
I led this high-visibility, ambiguous, zero-to-one initiative by partnering closely with product and engineering to define the problem space, establish clear desired outcomes, and guide solution discovery.
I led this high-visibility, ambiguous, zero-to-one initiative by partnering closely with product and engineering to define the problem space, establish clear desired outcomes, and guide solution discovery.
Vision and clarity
Developed visually engaging presentation decks to communicate vision and progress.
Rapid iteration
Built momentum through quick loops of ideation, validation, and iteration — always moving closer to what works best.
Empathy in action
Conducted targeted user research to uncover nuanced behaviours and validate opportunities.
Empathy in action
Framed design decisions — outlining tradeoffs, presenting clear recommendations, and connecting choices to business impact.
Vision and clarity
Developed visually engaging presentation decks to communicate vision and progress.
Rapid iteration
Built momentum through quick loops of ideation, validation, and iteration — always moving closer to what works best.
Empathy in action
Conducted targeted user research to uncover nuanced behaviours and validate opportunities.
Empathy in action
Framed design decisions — outlining tradeoffs, presenting clear recommendations, and connecting choices to business impact.
UX direction
UX direction
The proof of concept informed several design decisions, including moving away from a chat experience and the ability to add serving sizes.
The proof of concept informed several design decisions, including moving away from a chat experience and the ability to add serving sizes.
What's the scoop with serving sizes?
What's the scoop with serving sizes?
One of the biggest pain points with logging manually is that there are often too many (or not enough) serving sizes to choose from.
With the ability to articulate exactly how much you consumed (numerically and colloquially), voice log allows users to save time and frustration.
One of the biggest pain points with logging manually is that there are often too many (or not enough) serving sizes to choose from.
With the ability to articulate exactly how much you consumed (numerically and colloquially), voice log allows users to save time and frustration.
During testing, we discovered that the LLM can parse what a "sleeve of Oreos" is.
It's 13 Oreos.
Not sponsored!
During testing, we discovered that the LLM can parse what a "sleeve of Oreos" is.
It's 13 Oreos.
During testing, we discovered that the LLM can parse what a "sleeve of Oreos" is.
It's 13 Oreos.
Not sponsored!
See more case studies showcasing my design impact
See more case studies showcasing my design impact