Industry
Health & Wellness
Project date
2024
Platforms
iOS, Android
Opportunity
Help users quickly understand and get value from the new voice-based food search. How might we educate them effectively with minimal friction during first use?
Results
→ Converted more (query to log)
→ More logs containing multiple foods
→ Fewer errors encountered
↑ 3.4%
Voice query to
voice log CVR
↑ 3x
more likely to complete a
food log containing ≥2 foods
↓ 17%
Voice query
to error rate
Initial approach: do what we've always done
The first iteration of this feature was to lean into the tutorials we already used in-app. Aiming for consistency, I worked with my content designer to create a series of tips on how to use the feature.
Above: These lengthy flows would force users to tap through many screens before ever experiencing the feature. It’s like cramming for an exam the night before.
Applying my training from SVPG (silicone valley product group)
Instead, I used the product operating model (POM) to understand what users actually want to learn
How might we maximize the amount of education with minimal friction during a user’s first use?
Product operating model (POM)
Conducted interviews with users interacting with the voice log feature for the first time
Step one
Learnt about preferences for educational moments in app
Step two
Gained confidence around the type of content we should focus on and applied learnings to the new experience
Step three
View the research synthesis ->
(Summarized by AI)
Data analysis
Insights from user testing
0 out of 4 users knew they could refine their search using brands
2 out of 4 users said audio clips with examples would be helpful
Saying the brand is a power move that can help cut down time to log and provide more accurate results.
What are your thoughts on tutorials in apps?
What makes a tutorial helpful or unhelpful for you?
Interview question
4 out of 4 users mentioned that a way to skip or opt-out easily was important
🙅🏽
Insight
“Helpful: 1,2,3, lists…
Unhelpful: 10 screens showing how to use”
☑️
User quote
“Clear call to action what you can and can’t do”
💬
User quote
Visual explorations
After learning about the type of tips users want to learn about, I worked through a few visual explorations of how we could present the tips.
System composition
I designed a clear, scalable visual system that used a real voice log phrase as the anchor, revealing different highlighted elements on each tip to teach key features progressively while keeping the experience simple and unobtrusive.
Card Anatomy
→ Base UI layer: Recessed surface with button
↪ Content layer: Base tip content (anchor)
↪ Contextual layer: Highlighting key features
A scalable approach
By leaning into this visual approach, I explored how the coloured highlight treatment could scale into a systematic way to spotlight future enhancements.
think:
→ Feature announcements
→ Insights and stats
→ Streaks and engagement
What we shipped
This test supports the hypothesis that education tips help users use Voice Log more efficiently—driving higher query-to-log conversion, fewer errors and retries, and more foods added per log.
1st open
Serving sizes
2nd open
Multiple foods
3rd open
Brand names
