To comply with my non-disclosure agreement, I have omitted confidential information. The work presented has been shared with permission.
MyPalate is an early stage startup building a personalized recipe app to provide more cultural diversity and dietary accommodations. I worked on Search.
timeline
6 months Jun to Dec 2020
team
Christina 👋 Product Designer
Tancred, Derek  CEO & CBO
Thomas, Minh, Viet, Nihal Development
Calvin, LJA  Design
(and the rest of the MyPalate team!) 
background
It's hard to find recipes.
First briefly explain MyPalate mission and user goals, maybe short persona with visuals. Then another short paragraph bringing context of current app MVP and testing. The previous search experience was developed during a month-long sprint to develop an MVP of the entire app, so it was limited in functionality. User testing quotes visual.
Young professionals aged 20-30, with dietary restrictions
Our target users' goal is to cook culturally authentic recipes that meet their health needs & dietary preferences. They share frustrations of aimless Googling to find healthy alternatives for traditional recipes, taking time out of their busy schedules. 
Turnip greens yarrow ricebean rutabaga endive cauliflower sea lettuce kohlrabi amaranth water spinach avocado daikon napa cabbage asparagus winter purslane kale. Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic.
How Search worked, before I joined.
insights
Content is king.
Users expressed x. Categorization and kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic. Turnip greens yarrow ricebean rutabaga endive cauliflower sea lettuce kohlrabi amaranth water spinach avocado daikon napa cabbage asparagus winter purslane kale.
Searching for recipes is time consuming.
Explain why. Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic. Turnip greens yarrow ricebean rutabaga endive.
High dropoff on empty search result pages.
Explain why. Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic. Change order of this list depending on case study content and address the last 2 concerns in scope/feature chart (near flow). 
goals
Bring the search experience on par with competitors.
Explain why. Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic. 
Increase satisfaction of users looking for recipes.
Explain how this will be measured. Eg. rating, time spent? Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tomatillo melon azuki bean garlic. 
defining scope
Scope and this changed many times many iterations. Map out to focus on for this release. 4 parts. Can also talk about constraints here (technical and other). Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tomatillo melon azuki bean garlic.  
information architecture
Understanding mental models
Something about categorization, mental models, structure, maybe content strategy information this is the story of spreadsheets get ready. Also talk about how this affects the whole app not just search. Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic. Start with 2 key issues/goals
I used this info to x something inform conduct card sort list why goals and number of participants and recruitment methods and screening and survey questions. Talk more about methodology here. Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic. Maybe combine stuff so you have less tables. 
Turnip greens yarrow ricebean rutabaga endive cauliflower sea lettuce kohlrabi amaranth water spinach avocado daikon napa cabbage asparagus winter purslane kale. Veggies es bonus vobis, proinde vos postulo essum welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic.

Hybrid

Open

Found that in both x, were grouped similarly. endive cauliflower sea lettuce kohlrabi amaranth water spinach avocado daikon napa cabbage asparagus winter purslane kale. Veggies es bonus vobis, essum magis kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic.

Updated categorization structure

prototyping
Browse
The prior browsing screen contained a list of recipes, which looked too similar to other sections of the app, and had no way of organization. Working with a limited set of tags, I split the recipes by cuisine with plans of exploring other categories such as time, meal, and occasion in the future, as well as personalization. Talk about research process, constraints and what shaped design decisions
Explorations
Turnip greens yarrow ricebean rutabaga endive cauliflower sea lettuce kohlrabi amaranth water spinach avocado daikon napa cabbage asparagus winter purslane kale. Veggies es bonus vobis, proinde vos postulo essum magis kohlrabi welsh onion daikon amaranth tatsoi tomatillo melon azuki bean garlic. Add specs. Rerecord prototype to tap on a category. 

The starting point for searching

Suggestions
The previous version did not have query suggestions that populated as a user typed, but I noticed that other top recipe apps all had some form of it. I considered both first time and returning users, no input and filled search bar states, as well as disambiguation and priority of suggestions, to develop product requirements. Rerecord prototype to start from browse.

Suggestions for "Kimchi"

Show recipes
One of the main aspects that sets MyPalate apart from other apps is our focus on providing an experience tailored to people's unique dietary and health preferences through automatically showing suitable recipes. Because of this, the previous MVP displayed recipes in a way that showcased every dietary tag on each recipe card. At first, this made it really clear to users that they were being shown recipes that met their needs. But I hypothesized that this would feel like redundant information after spending some more time on the app. So, I designed a process to inform first time searchers.

First time searcher

Refine results
One of the main functionalities we were in dire need of was the ability to refine search results. Now, after analyzing several top recipe apps, I noticed there was a wide range of categories to filter and sort by – cuisine, time, ingredients, occasion, meal, and more, you get the point. As a small team, we could not add all these categories in one go. To decide the priority of categories to focus on, I relied on a mix of competitive analysis and surveying users on what information was most useful when looking for a recipe.

Finding recipes under 15m

No more matches
Being in early stages of development, the amount of recipes on our app was limited. We worried that this would affect user retention. People might leave our app if they can't find what they're looking for. So it was important to let the user know when they had reached the end, and give them some alternative options, to prevent dropoff. Surprisingly, many of our top competitors did not have great empty states – some even showed blank screens of nothingness. Our app already had a recipe request feature, which I implemented into the search flow. Depending on if filters were applied, I encouraged them to edit their criteria. And then below that, I showed them more recipes that could be a similar match. 

Can't find the one?

Oh, and I also had the opportunity to talk a bit about search, as well as give an intro to design processes in UX. I did this alongside Calvin at McMaster Designathon – an event with 200+ attendees, where MyPalate was invited as a sponsor. Over 48 hours, we guided designers with their projects and judged the final submissions. Anyways, if you'd like to read more about the process behind my design decisions, I'll be updating this page soon! 
What's next
Here I will talk about what's next
Back to Top