UX UI Case Study
2025
What is COOKD?
Cookd app helps people find easy step by step recipes, plan weekly meals, and add ingredients to a shopping list. It also suggests recipes from what you already have with a "Cook With What You Have" feature and offers DIY kit purchases in select locations. The app targets beginners and busy cooks who want quick meals.
Primary Research
I started by reviewing recent feedback from the Play Store, App Store, Reddit, X, and Instagram, gathering all insights related to search, filters, and discovery. Then, I reached out to active Cookd users on Instagram and LinkedIn for brief interviews.
Public Reviews
User Interview Insights (Filtered and Translated)
Dhanushiya, 27, working woman, regular Cookd user
"When I’m searching, I wish it showed ideas as I type. Right now I have to type everything and then hit search, which feels slow."
Jey, 29, fitness coach, Cookd user
"I’m a vegetarian, but there’s no quick veg filter in search. It’s hidden somewhere in my profile, and honestly, it’s no use there."
Mirdhu, 25, content creator, Cookd user
"I tried the filters when I first started using the app… now I don’t bother. I just scroll until I find something, because the current filters don’t really help."
Akhil, 32, PG Student, Cookd user
"I mostly use ‘Cook With What You Have’, but it’s buried inside the menu. Why not put it up front? It’s the most useful thing for me."
After the primary research, I was able to narrow down the core problem along with the sub-problems that needed to be addressed.
Core Problem
Cookd’s search and filter flow is slow and unintuitive, making it hard for users to quickly find suitable recipes. Limited filtering options add to user frustration and longer search times.
Sub Problems
Basic, Non-Assisted Search
No real-time suggestions while typing keywords (e.g., “chic” → “chicken biryani”).
No recent searches or recently cooked history for quick access.
No personalized suggestions shown before search, especially for first-time users.
No natural language or inclusive search support
Users can’t search in mixed languages, regional slang, phonetic spellings, or casual queries (e.g., “potato evng snacks”, “muttai poriyal”).
This excludes a large portion of regional and conversational users.
Limited and unintuitive filters
Only “Cuisine, Meal Time, Meal Courses” are available, key filters like dietary preferences, ingredients to avoid, nutrition, and cooking time are missing.
No quick way to clear multiple filters at once.
Filter options inside categories are not ordered by relevance or frequency of use.
Low discoverability of “Cook With What You Have” feature
One of the most loved features is buried inside menus, making it hard to find.
Users confuse it with general search because the differences are unclear.
Why Its Important?
A recent report shows that 61% of users will leave a content rich platform if they can't find what they're looking for within five seconds.
Effective search and filtering can drastically reduce bounce rates and increase engagement, as users are able to efficiently access relevant content.
Research indicates that 70% of users prefer using filters to refine search results, and content platforms offering advanced filtering options see up to 50% higher engagement rates.
Additionally, 81% of users report that thorough filtering options significantly enhance their content discovery and overall satisfaction.
Secondary Research
I usually place the current UI in a whiteboard tool, freely brainstorm ideas, and note down every thought without overthinking. I revisit and refine these notes over the next two to three days until I feel fully in sync with the product.
UI Audit & Ideation (Current Flow)
I then read articles on how filtering works in apps covering filter types for smaller screens, key guidelines and considerations, and ways to design a clutter-free search and filter experience.
Competitive analysis
I then moved on to competitive analysis. Since the Cookd app is mostly used by busy people and bachelors, they likely expect search and filter functions to work similarly to the apps they already use. So, I studied popular ones like SideChef, Cookpad, Tasty, Meal Planner (Samsung), and Ekilu, along with a few Indian food apps familiar to my target users. I broke down each app’s flow, examined how it worked, and noted every detail.
Competitive Analysis Board
Wireframes
Now comes my favorite part - paper and pen. No restrictions, no limits - just my space to sketch whatever comes to mind. For this case, after enough reading and analysis, I had three to four ideas ready, so I began sketching them out.
Concept Sketches
Final Designs
In the redesigned onboarding flow, I plan to gather essential user preferences such as the type of meals they usually cook, the time they typically spend cooking, their diet type, and any ingredient restrictions. Collecting this data early is crucial for delivering a truly personalized search and filter experience. It allows the app to surface more relevant recipes, save users time, and make the overall experience feel tailored from day one. Currently, the Cookd app doesn’t collect any of this information, which is a missed opportunity to connect with users right from the start.
New Onboarding Flow
Search Flow
Real-time Suggestions
As users type, the search bar will suggest matching recipes (e.g., typing “chic” shows Chicken Biryani, Chicken Grill, Chicken Tikka).
Shows Recent Searches and Recently Cooked recipes for quick access to past activity.
Relevant dishes appear directly below the search bar, allowing users to skip the listing page and start cooking faster.
New Search Flow
Personalized Suggestions
If there’s no search or cooking history, the space displays tailored recipe recommendations based on onboarding preferences.
Personalized Suggestion
Natural language and Inclusive search support
Implementing Natural Language Processing (NLP) in Cookd’s search will make it more inclusive, intuitive, and user-friendly by understanding the intent behind what users type regardless of spelling errors, mixed languages, abbreviations, or conversational phrasing. This ensures faster recipe discovery, reduces search frustration, and makes the app more accessible to a diverse audience, ultimately improving engagement and user satisfaction.
Covering multiple types of searches Like:
Potato evng snacks — (potato evening snacks) [abbreviation / short form]
Muttai poriyal recipe — (egg stir-fry recipe) [Tamil + English mix]
Paner butter masala — (paneer butter masala) [spelling mistake / phonetic typing]
Something quick with bread — (quick recipes using bread) [natural language request]
Kids tiffin idea — (ideas for kids’ lunchbox) [Hinglish / mixed language]
Chawal ki side dish — (side dish for rice) [direct phrase / regional language mix]
Airfryer pakoda low oil — (low-oil pakoda in air fryer) [keyword stacking / intent focus]
Natural Language Processing (NLP) in Cookd’s search
Filter Flow
Always-On Filters
Placing essential filters like Ingredients, Dietary Preference, Cooking Time, Meal Time, Meal Course, Cuisine, Ingredients to Avoid, and Nutrition in a horizontal chip format ensures they’re always visible, easily accessible, and quick to apply without navigating away from the results. This layout allows users to refine their search in real-time with minimal effort, improves discoverability of filtering options, and reduces friction in finding the most relevant recipes.
Always-On Filters
Acts Independently
Each filter works on its own, reducing dependency on other selections and minimizing cognitive load.
Designed for users who prefer applying only one or two filters, avoiding complex multi-step filtering.
Complex search intent is already handled by NLP-powered search, making filters a quick refinement tool rather than a necessity.
The “Show Results” button updates in real time, giving clear feedback on how many recipes match the selection before applying it. Users immediately know the impact of their choice without guesswork, improving confidence and decision-making speed.
Independent Bottom Sheet Filters
Here come the Bucket Filters - pick what you want and hit search
Designed for people who like their dishes a certain way, the filter order adapts based on your onboarding answers. For example, if you’re vegetarian, “Dietary Preference” appears first.
All Filters Area
Users can search across all available filters, making it faster to find exactly what they need.
Chosen filters are pinned to the top, giving users quick access for review or removal.
The CTA button updates dynamically to show how many results match the current filters.
If no results are found, the CTA button instantly shows “0 Results,” avoiding the need to navigate back to check.
Users always know which filters are applied and how they’re impacting the search results.
Filters Behaviors
Empty States
I Replaced generic empty-state illustrations with a comic-style Cookd logo character that “travels” with the user through their journey, adding personality and relatability.
For moments when search or filters return no results, the character conveys the message in a lighthearted, empathetic way instead of a plain message.
Always display a set of dishes closely related to the user’s last search or applied filters, ensuring the page is never completely blank.
Related suggestions gently guide users to alternative options, increasing the chance they’ll choose another dish instead of abandoning the search.
Keeps the user engaged even in dead ends, reducing frustration and increasing time spent on the app.
Empty States
"Cook With What You Have" flow
“Cook With What You Have” is one of Cookd’s most loved features for suggesting recipes based on available ingredients. However, it’s hidden in the app, and some users confuse it with general search. While the general filter also allows selecting ingredients, I designed a new, clearer flow for this feature.
A prominent home page card with a clear click area. To make it engaging and align with Cookd’s unique, out-of-the-box marketing style, I added our favorite Vadivelu empty pocket sticker. This not only explains the feature in a fun way but also boosts brand recall, making it more memorable for users.
- Quick ingredient sets (Popular Combos) made for typical “dry day” situations, so users can still find recipes even with limited supplies.
Users can select ingredients manually or search to find and pick desired items.
The CTA shows how many dishes are available based on selected ingredients.
If no dishes match, users can adjust or reduce ingredients directly on the same screen to avoid back-and-forth.
Dish listings appear like the general search page with auto-selected ingredients, allowing users to easily tweak filters.
"Cook With What You Have" Flow 2
Usability Testing Summary
To evaluate the efficiency of the existing Cookd app’s search and filter flow versus my redesigned version, I conducted a comparative usability test. I asked users to complete the same task in both versions.
Participants first completed the task using one flow (either current or redesigned), then repeated the task using the other flow. During each attempt, I recorded time taken, number of interactions, and any difficulties encountered. After both tasks, I gathered user feedback on ease of use, satisfaction, and preferences.
A total of 25 participants were tested - 80% Cookd’s regular users and 20% random users unfamiliar with the app.
Old Flow: Average task completion time – 54 seconds.
New Flow: Average task completion time – 36 seconds
The redesign achieved a 33% reduction in task completion time, demonstrating a significant improvement in search and filtering efficiency.
Challenges & Learnings
Recruiting Users for Interviews
Challenge: Finding users across platforms was extremely time-consuming - no replies, sudden drop-offs, “are you a scammer?” doubts, and many rejections. This ate up the largest chunk of my project time.
Learning: Never compromise on user interviews. The conversations revealed insights I never expected, My own perception of the app turned out to be very different from actual user behavior. Worth every minute.
Stuck During Brainstorming
Challenge: Midway through ideation, I felt unmotivated, struggled to nail the perfection I wanted, and at times felt disconnected from the process.
Learning: Staying consistent with my work made a big difference, and taking short breaks helped me return with fresh energy and ideas.
Technical & Design Constraints
Challenge: Understanding how to simplify filter groupings without losing advanced options was tricky. Mapping old flows to new ones revealed hidden dependencies in the UI, and redesigning without breaking the logic took multiple iterations.
Learning: Always design with system constraints in mind early on - it saves a lot of back-and-forth during prototyping.
Handling Multiple Design Directions
Challenge: I had dozens of visual directions for the UI. The temptation to explore outside my planned scope was high, but it risked derailing my timeline.
Learning: Staying within a clear design boundary kept me aligned with my research goals and helped me finish without major detours.
The END
Every product has room for improvement and the deeper you dig, the more hidden opportunities you’ll uncover. Testing with real users, iterating with focus, and removing friction not only makes their lives easier but also drives real business growth. Keep things simple, don’t overcomplicate, and trust your instincts. The best solutions often come when you’re enjoying the process. Stick with what feels right, keep learning, and design with curiosity. Cheers🥂!
Other Work
UI Showcase