In this case study, I’ve presented my work on Partnr and outlined my full process as clearly and concisely as possible.

🌟 Project Overview

What is Partnr

Partnr is a subsidiary of Rockman Group (Rockman Industries is an auto component manufacturer and part of the Hero Group). It is a quick-commerce venture providing spare parts for two-wheeler Mechanics and garages.

Project Type

A UX redesign aimed at simplifying the part search and purchase experience within a mobile app.

Business Goal

To increase the number of orders and reduce failed searches by making the app more user-friendly, ultimately driving an increase in the company’s GMV (Gross Merchandise Value).

Duration

This was a 2–3 month project.

Team

I worked alongside the CTO, 6 engineers, and 2 PMs from BCG. Together, we built a system that successfully served over 10,000 mechanics.

🌟 My Role

Title

I was the Lead Product Designer on this project.

Responsibility

I led and executed all field research, including in-person interviews with mechanics in Delhi and Bangalore. Streamlined the design process, purchase experience and established a scalable design system.

Worked With and How

I worked with the sales team to gain access to real-world mechanic shops, and synced with the CTO and PMs to align user needs with business goals.

Key Design Challenge

I needed to simplify the part search experience for users who were often uncomfortable typing in traditional search bars or lacked the time to scroll through a catalog. The goal was to make their purchase flow seamlessly fast and exceptionally smooth.

💔 Problem & Background

Core User Problems

1. Mechanics couldn’t find the Partnr app icon on their home screens.
2. Mechanics were not able to quickly find the parts they needed on the app.
3. Mechanics preferred placing orders over phone calls instead of using the app.

Problem Discovery

Through user interviews and shadowing, I discovered that many mechanics struggled to recognize the Partnr app icon among other apps when asked to open it.

They preferred to order through calls during work hours rather than open app and place order

Additionally, most mechanics didn’t need access to a wide range of parts — they primarily searched for high-demand items relevant to their work. For example, a mechanic who mostly repairs brake systems mainly looks for brake shoes.

☄️ My Approach

Understanding The Problem

I conducted around 30 in-depth interviews, spent 9–10 days observing mechanics in their shops, and mapped out the journey of how they currently order parts.

Ideate and Test Solutions

We divided the defined problems into two categories: quick-to-implement solutions and those that would take more time.

For the quick wins, we created final designs, launched them, and tracked their performance using Mixpanel while also gathering user feedback.

In parallel, we began developing and testing the more time-consuming solutions, using insights from the initial rollouts to guide the next steps.

❤️‍🔥 The End Result

Key Design Decisions

☝️As I have mentioned, One major issue was that Mechanics couldn’t find the Partnr app icon on their cluttered home screens. We tried multiple solutions — changing the app icon, sending notifications — but nothing worked.

Then came a new idea: What if we made the app icon as big as possible? — That’s how the widget came in.

Widget displayed a rotating list of spare parts with prices directly on their home screen, making it impossible to miss.

This approach improved retention by 24% and GMV by 4.8%

☝️ Another major gap was that, as a company, we were selling products that — we wanted to sell (to make a profit) rather than the ones Mechanics actually wanted to buy.

To fix this, we:

✅ Allowed मैकेनिक्स to select their top 4–5 most profitable parts
✅ Introduced voice search & image upload to precisely capture their requirements
✅ And Customized the homepage to match their needs

And this personalized approach increased — add-to-cart rate by 34.6% and GMV by 10.8%.

☝️ I found that Mechanics mostly repair one or two specific two-wheeler brands based on their area’s demand. For example, if a mechanic primarily works on Hero Splendor, they might need related parts frequently and would place an order over a call.

Now, the challenge was: how do we make ordering as fast and effortless as calling a supplier?

To solve this, we:

✅ Created custom bike specific Bar code sheets
✅ Pasted these sheets on workshop walls for instant access
✅ Mechanics could simply scan barcodes via the app to add parts to their cart and place orders in seconds

This solution led to a 27.2% reduction in order time and an 18.9% boost in GMV.

What changed for the user?

Mechanics were now able to order the parts they needed without hassle, confusion, or wasting time on the app.

🦭 Learnings & Reflections

Learn About User

I learned that even the simplest UI can fail if it’s not rooted in the user’s world — literacy, habits, and expectations matter more than screens.

Project Shaped Me As A Designer

It made me more confident in designing beyond assumptions — especially for users who are rarely represented in mainstream UX patterns. Going forward, I want to bring this low-tech, human-centered mindset into every project I do.

Partnr’s stakeholders wanted to roll out exciting offers and deals through banners. What they actually rolled out? Pure, unfiltered mediocrity.

Story Time 📖✨

Partnr, a QUICK-commerce venture supplying spare parts to two-wheeler mechanics and garages, QUICKLY decided to introduce banners and QUICKLY failed.

Before this experiment, they knew nothing about banners — but after it, they were at risk of knowing less than nothing.

But they realized they didn’t hire me just to show up on time and leave — never.

THEY WANTED ME TO BRING BANNERS BACK FROM THE DEAD

Run, Forrest, Run! 🏃‍♂️

I started my work and found out that 2nd, 3rd, and 4th banners had extremely low click rates, preventing the company from effectively promoting its deals and offers.

I felt the need, the need for User Research

Peekaboo! 👀🙈

During user research, mechanics didn’t give a “single nut 🔩” about the banners — or me.

But resilience prevailed, one thing I learned from Bollywood: how to watch your p̶r̶e̶y̶ users silently — without them ever knowing.

This technique prevents them from the Hawthorne Effect. BIG WORD 😏

The Hawthorne effect means people change their behavior when they know they’re being watched.

I found that mechanics mostly use apps with vertical scrolling, like YouTube Shorts and WhatsApp. They weren’t comfortable with horizontal scrolling.

Magic 🪄

I finalized three designs.

Design A : I took inspiration (not copying… or maybe 👀) from other apps’ auto-scrolling banners.

Design B : I wanted mechanics to see all banners at once, so I grouped them together.

Design C, I took inspiration from the legendary telephone directory — like how it made switching and finding a number in a sea of names easy.

I tried the same in my design making it easy for users to switch between banners.

Judgment Day 🔥

Design C boosted mechanic engagement, leading to a 51% relative increase and moved banner out of the “just there, but never clicked” zone.

Takeaways ✅

Design A — Auto-scroll banners were like a suspense movie dropping clues over time but with humans having an attention span shorter than a goldfish, nobody stuck around for the plot twist.

Design B — I grouped the banners together like a family but turns out dysfunctional, making them hard to read/understand.

Design C — Placing banner previews on the right side gave users full control. Visibility went up, chaos went down.

Clap!! Comment!! Share!! & Have Fun

Joining a company as a Product Designer often involves walking a fine line between meeting stakeholder demands and addressing user needs. Stakeholders may sometimes request screens or features without fully considering their impact on user experience.

Here’s the story of how I advocated for the user while maintaining professionalism and trust.

🏢 Company: Partnr
👥 Users: Mechanics

What is Partnr?

Partnr is a quick-commerce spare parts venture designed specifically to cater to two-wheeler mechanics and garages. It provides a platform where users can:

1. 🛍️ Browse products from a vast inventory.
2. 🛒 Select and order required spare parts.
3. 🚚 Receive same-day delivery, ensuring mechanics can complete repairs promptly.

How Does Partnr Ensure Same-Day Delivery? 🚚

To maintain its same-day delivery promise, Partnr operates through a network of small dark stores strategically located to cover service areas efficiently.📍🏪

Additionally, Boston Consulting Group (BCG) has been instrumental in setting up this ecosystem, ensuring operational and logistical excellence.

Target Audience 🎯

Partnr’s audience is highly niche, consisting primarily of:

• Two-wheeler mechanics
• Garage owners

Due to the specificity of this audience, online advertising alone isn’t enough to reach potential users effectively. 📡

Sales-Driven Outreach Strategy 🔧

To overcome challenges in reaching this unique audience:

• Partnr relies on an on-ground sales team to directly interact with mechanics.
• The team educates them about Partnr’s offerings and helps them download and use the app.

This personalized approach ensures mechanics understand the value proposition and are onboarded smoothly, reducing the dependency on time-consuming digital ad campaigns.

Partnr bridges the gap between mechanics and spare parts with speed, efficiency, and a user-centric approach! ⚙️

Joining Partnr

When I joined Partnr in February, the venture was just starting its services in Delhi. At that stage, the company relied on a Shopify-powered website to facilitate initial orders.

Challenges with Shopify Website

• 💸 Cost Management: Running operations on Shopify came with recurring expenses that didn’t align with their long-term goals.
• 🛠️ Full Control: Partnr wanted to build a platform entirely tailored to their business model, ensuring complete control over features, design, and updates.

🌟 My Role

I was brought on board as the Lead Product Designer to help them transition from Shopify to a custom app. My responsibilities included:

• 🖌️ Designing a scalable app tailored to Partnr’s unique business and user needs.
• 🎯 Creating a seamless, user-friendly experience for mechanics and garages.
• 🛠️ Helping Partnr take ownership of their platform, ensuring flexibility and cost-effectiveness in the long run.

This shift was pivotal for Partnr, allowing them to break free from third-party dependencies and build a product designed for their vision and audience. 💪

A Challenging Start: Designing Without User Research

As soon as I joined, I was handed the responsibility to design the entire app. But there was a twist:

• ❌ No user research had been conducted yet.
• 📊 Limited resources: The only data available was the feedback from the sales team.

💡 My Approach: Start Small, Grow Big

My design philosophy is simple “Design the app to meet current user needs, rather than cluttering it with future-ready but unnecessary features.”

📌 The Plan in Action

1️⃣ Focus on the Essentials:

• Initially, I designed only the features that were immediately relevant to users.
• No distractions: Avoided adding fancy or irrelevant features that wouldn’t be necessary in the next 3–4 months.

2️⃣ Data-Driven Simplicity:

• Leveraged available data: Analyzed insights from Shopify and feedback from the sales team to prioritize features.
• Focused on creating an intuitive and seamless user experience.

3️⃣ Iterative Versions:

• Created multiple simple and user-centric versions to ensure users wouldn’t be overwhelmed by unnecessary options.
• Left room for future enhancements based on user feedback and evolving needs.

This lean and focused approach ensures the app is relevant and user-friendly for current users while keeping it flexible for future growth. 🌟

Deciding the Homepage Structure 🏠

After analyzing all the available data, I finalized the homepage layout with key elements that balanced user needs and stakeholder priorities. The homepage prominently displayed part categories to ensure mechanics could quickly find relevant parts, addressing their specific needs. I prioritized audio search as the primary focus, based on my intuition and past experience that mechanics often avoid typing due to convenience or discomfort with text input. I believed voice search would be the simplest and fastest way for them to reach their desired products, making it crucial to start development early as a potential game-changer for user experience. To align with stakeholders, I also included a traditional text search 🔍, ensuring users who preferred typing had the option while reassuring stakeholders about the design’s completeness.

To create a seamless experience for mechanics using the app, I focused on making the homepage design clean and user-friendly. I intentionally avoided clutter by not cramming all categories into a small area; instead, I ensured category images were larger, making them easily recognizable and visually engaging. The layout was designed to feel spacious and not overwhelming. I created a design that encouraged natural exploration, rather than forcing all categories into the first fold.

For the filter feature, I thought a lot about how to make it more intuitive and easily recognizable for mechanics. I decided to use brand logos instead of just text for two reasons.
First, quick recognition is key, as mechanics often deal with specific brands, and logos are visually familiar to them. By using logos reduces cognitive load because mechanics don’t need to read through names; they simply look for familiar logos, making the filtering process faster.
Second, enhanced visual appeal 🎨 was achieved by incorporating logos, which not only make the filter more engaging but also improve the overall aesthetics of the app.

🚧 The Stakeholder Challenge: Gut Feeling vs Data-Driven Design 🤔

Despite my efforts to align stakeholders with the “start small, grow big” approach, I couldn’t bring them on the same page. The stakeholders firmly believed that mechanics would prefer a full-featured app from the start and that launching a robust app with all the bells and whistles was necessary to capture the market.

I presented data showing that mechanics were often overwhelmed by too many features and argued that a lean, focused app would be more effective and easier for them to adopt. However, the stakeholders chose to rely on their intuition, believing that rolling out a full-featured app was the right decision.

While I disagreed, I recognized the value of following gut feelings; ❌ if the approach failed, it would be a learning opportunity, and ✅ if it succeeded, it would validate an instinct-driven strategy. As a designer, I learned that sometimes you need to let the process unfold, allowing others the space to experiment and learn from the outcomes.

🚀 Designing & Launching the App: Speed Meets Strategy

So, I started working on the app, creating all the essential pages including Onboarding, Home Page, Search, Filter, and Checkout. When there’s pressure to design quickly, the smartest approach is to always have a basic design system ready. This allows me to reuse common elements across the app, making it easier to maintain consistency and save time in high-pressure situations. ⏳✨

Even in such conditions, it’s crucial for me to keep the user in focus. No matter if the design comes to life from someone else’s gut feeling, as a designer, my responsibility is to execute with full conviction, not half-heartedly.💪

To address stakeholder expectations while maintaining a user-centric design, I combined their requirements with UX-friendly solutions.

They wanted to include banners, which I was strongly against because, in my experience, the opening rate for banners is usually very poor. Moreover, they take up a lot of space unnecessarily. However, since they were insistent on adding banners, I came up with a way to improve their visibility and click rate. The solution was to display a preview of all banners on the right side, while the selected banner would appear on the left side in full view. This ensured better engagement without compromising too much space.

On the homepage, they wanted to showcase all categories along with the fast-selling products that drive the most sales. To avoid the design feeling monotonous, my idea was to display the main categories at the top, followed by fast-selling products styled differently, and then the remaining categories below. For the category layout, I used a 3-grid view to ensure that the categories didn’t look cramped and that the images appeared larger and more visually appealing. 🌟

I prefer designing everything with a vertical scroll because horizontal scrolling isn’t a common behavior among users. Moreover, most popular apps that shape user habits also prioritize vertical scrolling, making it a more intuitive and user-friendly choice.

Similarly, for the PLP (Product Listing Page) cards, the main requirements were to display all compatible vehicles for each spare part, along with an product image, a CTA (Call to Action), price, and the full product name. To ensure a better UX, I prioritized three key elements: the image, the CTA, and the price, ensuring they were distinct and clearly visible to the user. I also wanted to include the maximum number of compatible vehicles directly on the card, with other details arranged around these primary elements. As a result, I designed a card that balanced functionality and clarity seamlessly.

For the first time, I faced a new challenge where stakeholders insisted on shaping the UX and UI entirely according to their preferences. It seemed they were heavily inspired by another app or product. However, when you simply copy-paste ideas from other apps or products without adapting them, you’re unlikely to succeed. 🛑 You need to tweak the design, sometimes a little or significantly, to align with your user needs.

Unfortunately, their proposed UX was quite generic, and they ended up overturning most of my design decisions. This situation highlighted the difficulty of balancing stakeholder expectations with user-centric design principles.

They started by insisting on changing the background color, and then moved on to other aspects like making the banners horizontal. They wanted fast-selling products to also appear in a horizontal scroll, followed by all categories displayed in a 4-grid view with smaller-sized images.

The same approach was applied to the PLP cards, where their preference was to show the maximum number of cards in the first fold. This approach clearly prioritized quantity over clarity and usability, challenging the idea of keeping the design user-friendly and visually balanced.

Finally, I completed the entire app design and handed it over while the frontend and backend teams worked simultaneously on development. Together, we successfully launched the app in record time. 🚀

Post-launch, we guided our existing users, who had been using the website for orders, to download the app, and we also began onboarding new users.

💔 The App's Struggle

As expected, the app performed worse than the website. Despite significant efforts from the sales team, mechanics were either not using or unable to use the app effectively.

The design lacked heart and soul ❤️, which gave it a limited lifespan. As a designer, I could sense during development that it wouldn’t work, but in a professional environment, I often had to build what was asked, even if I knew it was flawed. Such designs could only be discarded when data 📊 or feedback 💬 provided irrefutable evidence of their failure. 🚫

However, the stakeholders were still unwilling to back down. They believed that by adding extra features and making some changes to the current designs, the app could be made successful. I was also instructed to create more designs.

This time, I decided not to shoot in the dark by creating random designs. This approach would waste ⏳ time, show minimal 📉 impact, and leave us with zero 🔍 meaningful learning.

To bring the stakeholders on board, I leveraged my past experiences, interactions with mechanics, and my understanding of how users in this tier perceive and use technology and apps. I conducted a UX audit of the current designs and submitted a detailed report to the stakeholders to convince them.

🎯 My Goal:
Through the audit report, I aimed to show:

• Why all future designs should be research-driven and thoughtful.
• Why I needed time for user research to back my designs.

As expected, the stakeholders repeated their stance “We don’t need user research right now.” They believed we should simply copy the user flows from apps like Boodmo and KhataBook, assuming these companies had already conducted extensive user research to arrive at their designs.

👊 Counterargument
To counter this, I sought help from my friends, Ankit Dua (Frontend Lead) and Rohan Dixit (Backend Team). My argument was:

• How can we be sure these companies have even conducted user research, let alone in-depth research? 🤷‍♂️
• We don’t have any 📄 research papers or 📊 data from these companies to validate this assumption.
• While copying their designs might sound easy, when you factor in the time cost of 🎨 designing, 🧑‍💻 frontend development, and 🔧 backend development, it becomes clear that we are working blindly. If the results don’t align, reversing these changes will be even harder and more costly.

I emphasized that starting with research is crucial, as it would at least provide us with a clear direction. Even if we spent two to three weeks on research, it would be worth it if it helped us work in the right direction.

Finally, I managed to secure permission for one week of user research. 🕐

🚀 Preparation Before Starting User Research

Before diving into research, I made sure to prepare thoroughly to maximize the limited time I had. Here’s what I focused on:

Business Goals:

• Understand how mechanics interact with our app.
• Identify gaps that are preventing app adoption.
• Ensure findings can directly increase adoption.

User Goals:

• Mechanics should feel the app is easy to use.
• Enable them to quickly find and order the spare parts they need.
• Address any confusion or hesitation they face with technology.

Given the time constraint, I ensured that my questions were direct — getting to the root of the problem — and actionable — helping me identify specific changes to improve the user experience. I focused on key areas such as:

• 🛠️ Onboarding: Was it simple? What hurdles did they face?
• 🧭 Navigation: Could they find what they were looking for?
• 🛑 Features: Which features were confusing or unnecessary?
• 🤝 Trust: Did they feel confident using the app to place orders?

📍 Planning the Research Flow
With time being critical, I prioritized:

• 👥 Identifying Key Participants: Targeting mechanics who had interacted with the app or were potential users.
• 📍 Choosing Locations: Focusing on areas where mechanics were most active.
• 🗣️ Tailored Interviews: Preparing for both in-depth discussions and quick observations based on the mechanics’ availability.

🎯 Goal of Research
The ultimate aim was to deliver data-backed insights to stakeholders, propose practical and impactful design changes, and gain clarity on why the current app wasn’t working and how to fix it.

This preparation ensured I could maximize the short time allocated for research and provide solid, actionable outcomes to guide the next steps. 🚀

📉 Failed User Research Experience

My first attempt at user research in Bhimnagar turned out to be a valuable learning experience, despite not achieving the results I had hoped for. I coordinated with the sales team to identify areas where we had the highest order volumes. Their suggestion led me to Bhimnagar, where I spent two days conducting user interviews with 22 mechanics. However, several issues arose during this research.

It became clear that the mechanics were not engaging honestly 🛠️. They provided no real insights about the app because the sales team handled everything manually for them. Moreover, there was a lazy attitude toward the app 🤷‍♂️. Since the sales team directly took orders to meet their daily targets, the mechanics had no motivation to use the app and were content with the manual process. Most of them didn’t even know how the app worked 🤔, which made the research ineffective.

This experience highlighted that I was talking to the wrong users. These mechanics had no incentive to use the app because the sales team bypassed it entirely. I realized I needed to pivot my approach ⚡ by targeting mechanics outside the sales team’s influence — those who had no prior exposure to Partnr’s sales team and might genuinely be interested in online spare parts ordering. Additionally, it became important to assess whether these mechanics were even open to adopting technology for ordering 📱, observing their attitudes and pain points to determine their readiness for change.

Although this was a setback, it helped me refine my research focus. It reinforced the importance of finding unbiased participants and digging deeper to uncover the root of the problem, ensuring future research would be more impactful. 💡

🔍 Another attempt of User Research

After the failed research in Bhimnagar, I decided to shift my focus to Badarpur, where I had some local connections. This time, I was determined to gather meaningful insights.

I began by leveraging my local contacts, reaching out to mechanics through friends who owned two-wheelers in the area. This gave me a starting point to initiate conversations with a small group of mechanics. From there, I expanded my network by building trust one mechanic at a time 🤝, eventually managing to interview 30 mechanics over three days.

Time management proved to be a significant challenge ⏳. With only one week of permission for user research and two days already lost due to the Bhimnagar attempt, I was left with just five days. Realizing I’d need at least one or two days to compile my findings 📋, I had only three days to conduct interviews and gather data.

Despite the tight timeline, this approach was far more productive. Mechanics in Badarpur were more open and honest, and my local connections helped break the ice quickly. However, the pressure to ensure every interaction counted was immense.

This experience underscored the importance of targeting the right audience 🎯 and managing time efficiently in user research, especially under constrained timelines.

✨ Key Learnings from Badarpur Research

Visiting Badarpur turned out to be a game-changer. It gave me confidence in some of my assumptions and helped refine the direction for new designs. Here’s how I turned these findings into actionable insights for new design directions:

🛠️ Learning 1: Focus on Profitable Spare Parts

• Observation:
  Every mechanic has a few specific spare parts that bring them the most profit, depending on the type of two-wheelers they primarily service. For example, If most of a mechanic’s business comes from repairing scooters (specifically brake shoes), then brake shoes become their key product.
• Missed Opportunity:
  Currently, the app is not selling what mechanics need; instead, it focuses on what we want to sell. This misalignment creates a gap between the user’s needs and our offerings.
• Actionable Insight:
  When new users are onboarded, gather immediate information about their profitable parts and showcase those parts on the homepage. Similarly, for existing users, push this feature as a notification when they open the app. Gather their input too and show them the parts that bring them the most profit on the homepage. This feature ensures that mechanics will at least buy what they already need. Other spare parts can be organized into a separate tab for easy access.

By implementing this approach, we can create a more user-centric experience while ensuring mechanics are more engaged with the app.

🔍 Learning 2: Fear of Duplicate Products

• Observation:
  Mechanics prefer to physically check spare parts upon receiving them to avoid duplicate or damaged products. If a product is faulty, they return it immediately.
• Missed Opportunity:
  The app currently lacks any messaging that assures users they can easily return faulty products. Additionally, the photo quality of spare parts on the app is subpar, which fails to build confidence in the product.
• Actionable Insight:
  1️⃣ Start by capturing high-resolution photos 📸 of the spare parts that bring maximum profit to mechanics and upload them to the app. Monitor whether this improves sales for those products.
  2️⃣ For long-term improvement, consider investing in a 3D viewing option 🎥, allowing mechanics to rotate and examine spare parts in detail before purchase.
  3️⃣ Add a guarantee badge 🏅 or a clear return policy message for all products, reassuring users about quality and returns.

These steps can build trust and improve the overall user experience, leading to higher engagement and increased sales.

🏪 Learning 3: Existing Sourcing Habits

• Observation:
  Most mechanics source their spare parts either from local dealers or by traveling to wholesale markets. For urgent needs, mechanics either call suppliers or send photos of parts on WhatsApp.
• Missed Opportunity:
  The app currently doesn’t allow mechanics to view their nearest dark store, nor does it provide a system to handle custom requests beyond the basic “add to cart and order” functionality.
• Actionable Insight:
  1️⃣ Enable mechanics to view their nearest dark store location 📍 on the app. This feature would allow them to visit the store, check the parts physically, and build trust before fully adopting the app. Over time, highlight the convenience and cost-effectiveness of ordering directly through the app to increase adoption rates.
  2️⃣ Introduce a feature where mechanics can upload photos 📸 of the required spare parts directly in the app for instant identification and ordering. This would mimic their familiar WhatsApp-based process while integrating it into the app ecosystem, making the transition smoother and more effective.

📲 Learning 4: App Usage Behavior

• Observation:
  Mechanics are far more comfortable using YouTube and WhatsApp compared to other digital platforms. On WhatsApp, they are proficient in attaching files, posting stories, and sending voice messages. On YouTube, they rely heavily on audio search 🎙️, avoiding typing wherever possible.
• Missed Opportunity:
  The app currently lacks audio search functionality 🎤, offering only a typed search option. Additionally, when mechanics share a product link, the recipient must download the app and complete the onboarding process before viewing the product.
• Actionable Insight:
  1️⃣ Introduce audio-based search in the app to align with mechanics’ existing habits and make the process effortless.
• 2️⃣ Enhance the share functionality by allowing mechanics to share product links as webpages 🌐 on WhatsApp. This way, the recipient can view the product’s photos, details, and price directly without downloading the app, making the experience seamless and increasing visibility.

All these learnings formed the foundation of the app’s new designs, ensuring they were aligned with mechanics’ real-world behaviors and needs. The goal had been to bridge the gap between what stakeholders wanted and what users truly required.

What passed and what not?

I shared many ideas, and one of the best strategies I learned for presenting your best ideas is to always include some weird or seemingly useless ones. This approach ensures that the decision-maker, who has the power to say “yes” or “no,” sees how much you’ve explored. It also helps them make better decisions when they can compare bad and good ideas side by side.

Not all my ideas were approved, but a few key ones made it through. For instance:
✅ We started collecting user preferences right after onboarding and introduced a personalized homepage.
✅ Audio search for spare parts was accepted.
✅ A feature to allow mechanics to send requests in a structured format directly from the app also got the green light.

These approvals felt like small wins, paving the way for a more user-centric app.

🎨 Implementing New Designs

Simplifying Product Search

To ensure that mechanics could easily find products from a vast catalog, the first priority was implementing audio search. Initially, my vision was to make audio search the sole option on the app, but the top leadership insisted on keeping text search as it was. This led me to rethink the implementation. My first thought was to place audio search as a floating action button (FAB), but I rejected this idea and instead adopted a YouTube-style implementation, ensuring familiarity for users by mirroring how they already interact with YouTube.

When comparing type search to audio search, we observed that audio search had 35% higher engagement. However, in the early stages, result accuracy was a challenge, as the model needed further training. Despite this, the progress signaled that we were moving in the right direction and that accuracy would improve over time.

Personalized Homepage

To ensure mechanics saw highly profitable and frequently used parts directly on their homepage, we began working on a personalized homepage experience. This required identifying parts they needed often by combining their frequent requirements with purchase history (if available).

For new users, we collected this data immediately after onboarding, while for existing users, we prompted them to provide information when they reopened the app. Although we tested multiple versions, all had a high closing rate. However, achieving the best personalization required rich data.

To gather this, we implemented audio input, allowing mechanics to speak their required parts, which our backend processed to personalize their homepage. To further improve accuracy, we added photo-based input, enabling users to upload part images, enhancing our data collection for better personalization. While our audio and image recognition models were not initially advanced enough to instantly recognize inputs and update the homepage in real time, they were essential for training and improving accuracy. In the meantime, we displayed all categories on the homepage during personalization processing.

Visibility and Usability Challenges in FSP Cards

The business decision was to keep Fast Selling Products (FSP) in horizontal scroll format, as mandated by stakeholders, to focus on products with high margins for the company. However, the major concern I identified was low visibility — mechanics rarely engage with horizontal scrolling, leaving many cards unseen.

The straightforward solution was to introduce a “View All” CTA that directed users to a dedicated page where all cards were displayed in a vertical scroll. Initially, the CTA followed standard practices, but its click rate was low. To address this, I redesigned the CTA, making it more prominent and placing it on the left side. This change resulted in a 30% increase in click rate, significantly improving engagement.

To further enhance the FSP card UX, I updated its UI:

1. Increased the space for product images, as mechanics heavily rely on visuals.
2. Added a preview feature for additional images, allowing users to view up to three more images in an enlarged format upon clicking.
3. Enlarged the price and slashed price text for better visibility.
4. Introduced a quick add CTA, enabling mechanics to add products directly from the card instead of navigating to the Product Display Page (PDP).

Category Visibility

The challenge was to decide how to display categories, as the initial design had very small image sizes, which was problematic. Mechanics rely on visual recognition, not text, to identify parts. Additionally, until the personalized homepage was implemented, all categories needed to be displayed on the homepage.

To improve the user experience, I redesigned the layout to show categories in a 2x2 grid. This ensured:

1. Larger images that were easier for mechanics to recognize.
2. Distinct visuals by assigning different colors to the cards, preventing them from blending together.

Once the personalized homepage was ready, we further optimized the design by removing text entirely from category cards, displaying only images to align with the mechanics’ behavior of relying on visuals. This approach made the interface both clean and effective.

Making the Homepage Vibrant and Memorable

To provide an even better experience, I focused on improving the top part of the homepage. Initially, it felt dull, so I decided to make it vibrant and relaxing. A good color combination is essential as it not only attracts users but also makes them feel good, whether they are from tier 3 or tier 1.

Previously, when a mechanic visited the app, they could see the branding, but it wasn’t memorable. To make it stand out, I added a tagline, ensuring that the current design had a more engaging and impactful message.

These small changes, while they may not drastically improve metrics, create a lasting impression on users. It’s these little details that users remember the most, and ultimately, user experience matters far more than any metric.

👑 How I Evolved

After completing this entire process, I realized how important it is to tackle the situation when top leadership gets heavily involved. It’s crucial to understand that you can’t always balance both business needs and user experience; but in the end user experience and its needs should always take priority. Business needs should emerge from user understanding, especially if you’re playing the long-term game.

Questioning top leadership can be tough, so it’s best to present your ideas, but don’t take things personally. There’s a chance your idea may never be implemented, but as a product designer, your job is to transform any idea, even a bad one, into the best user experience. You should take on these head-on challenges. Because within the company, disagreements will come and go, but if you lose the user, they may never come back.

🙏 Thank You

Throughout this entire journey, I am immensely thankful to three people:

Rohan Dixit (Backend Engineer) — I was amazed to see a backend developer dive so deeply into the UI, questioning and challenging every detail.

Ankit Dua (Frontend Lead) — Whenever I shared any idea, he always responded with the same confidence, saying, “It will be done.”

Ishaan Rajora (Frontend Engineer) — Even though he’s much junior to me, he taught me a lot about frontend development and helped me grow in that area.

🤟 Just Chill

If you’ve read through this long and detailed case study and have made it here, please comment so I can personally thank you. I have one more case study titled

How UX Became the Game-Changer for Funding

but don’t read it yet — I’m sure you won’t have the energy left. Take your time and give your views after you’re done relaxing!

If a metric in a startup drops from 27% to 16.5% and no one understands the reason behind it and series B funding on the line, then you are in for a show.

This is the story of how, amidst chaos, trusting yourself can transform an important company metric from 16.5% to not just 27% but an impressive 39.3%

🏢 Company: Animall.in
👥 Users: Dairy Farmers
🗒️ Note: Designs translated into English (original designs were in Hindi). Please ignore the listing content, as it is only used as a placeholder.

Introduction

Animall is a digital platform designed to connect dairy farmers, enabling seamless buying and selling of livestock.

Imagine a small dairy farmer in rural Rajasthan, struggling to find a buyer for his prized cow. With Animall, such farmers can now effortlessly connect with potential buyers, breaking geographical barriers and creating a thriving marketplace.

One key metric for its success was the CWA.

🤔 What Is CWA?

When Buyers contact sellers using WhatsApp or call button, CWA counts 👇

However, the metric experienced a severe decline, threatening the platform’s usability and business impact.

⚠️ The Crisis: When Metrics Fail

One Saturday night, while I was sleeping, I got a call: ‘Hey, do you know what happened? CWA crashed. I was wide awake right away. 😳

The CWA rate had plummeted 📉 from 27% to 16.5%.

A decline in CWA means buyers were not contacting sellers using WhatsApp or the call button, which directly impacted transactions.

This metric was reviewed by the Sequoia team every two months, and our asking amount for Series B funding heavily depended on this metric.

“I had so many questions: How did we miss that? Why was no one monitoring that number? Why, and so many more whys. But their answers wouldn’t do any good.”

🔎 Initial Observations

As soon as the next day began, I had prepared several assumptions because I knew the pressure to solve this quickly was about to hit. I had even thought of multiple solutions.

My top choices were :

Visibility of Negotiable Prices for Buyers
One of the challenges was that buyers might not easily determine which sellers were open to negotiating prices. This could make it harder for buyers to engage with sellers who might be willing to make deals.

Improving Photo Visibility for Buyers
When buyers evaluate livestock, they focus on specific attributes like the udder (थन) and horns (सींग) to judge the animal’s price accurately. However, in the current listing design, only a single photo is prominently displayed, and additional photos are accessible only via Horizontal scrolling. This design might fail to communicate effectively to users that more images are available, or the horizontal scrolling behavior might not feel intuitive.

Then, along with the Product Manager and developers, we decided which assumptions we could quickly act on. Price slash and Improving Photo Visibility on listings were the top choices.

Now, you might be wondering how all this is happening without user research, especially when the first rule of UX is user research followed by blah blah blah…

But in reality — especially in startups — when things go south, you have to trust yourself, rely on your past research, your preconceived notions, and take a leap of faith. You can’t do user research for everything, and stakeholders won’t give you that much time either.

Implement it first, Check it later!!

🔄 Implementation & Testing

To ensure accuracy and measure impact, the PM strategically released the experiments in controlled buckets and the both experiments showed some positive results.

However, aligning stakeholders for the slash priced experiment was quite challenging because the startup wasn’t in a position to take a hit on prices. They were looking for a long-term solution in a very short amount of time.

We moved forward with the photo visibility experiment but did not implement in all the buckets.

🚀 Another Attempt

We needed more assumptions that we could trust to have a meaningful impact, but we were out of ideas. That’s when power users came to the rescue. 🙌

“Power users are your old, loyal customers with whom you stay in regular touch. They don’t waste your time, nor do you waste theirs. They directly share their problems with you.” 💬

I had 4 power users and from one such power user, I discovered how he wanted to buy a buffalo 🐃 but couldn’t find one now, even though he had purchased a buffalo from the platform before. 😟🔍

Another user asked, “Can I find a Bull on this app?” 🤔

🛠️ Assumption & Validation

Meanwhile, the PM (Chirag) was busy analyzing data and discovered a surprising trend: as the number of listings increased ⬆️, the CWA numbers decreased ⬇️.

In simple words: Buyers had stopped calling sellers even as livestock listings were growing.

We checked the pattern of listings displayed on the homepage 🏠 and came up with this assumption:

“The homepage keeps showing the latest listings. So, if four cow 🐄 listings are uploaded and the user sees only cows up to the second fold but wants a buffalo 🐃, they drop off.”

Another assumption emerged: “If someone wants an animal that is rarely sold on the app, its listing might not even appear for them, leading to drop-offs.”

The main question arose: Why aren’t buyers using filters to search for the animal they want?

We checked the filter’s click rate 🖱️, and the results made it clear that we had to focus on improving the filter.

The filter’s click rate was just 18.9%, and even fewer users — only 11.2% — completed their tasks after using it. 😓

We defined our Goal and Success Metric 🚀

✨ Old Is Not Always Gold

The old filter was inspired by Amazon/Myntra-style filters, but clearly, it wasn’t vibing with our users. 💔

⏳ Tick-Tock Tick Tock

We didn’t have much time ⏱️. We thought if we could improve the click rate on the homepage filter, it would also boost the completion rate deeper in the flow.

I brainstormed some UX improvements/solutions that were simple to implement and had a good chance of showing results .

For example:

• Highlighting the filter button in different ways
• Updating its UX copy with more engaging text

These small tweaks aimed to make a big impact! 🚀

🚧 Dead End

Experiments like Iteration 1 and 2 failed to deliver the desired results but Iteration 3 boosted the homepage filter click rate— it improved from 18.9% to 31.2%. However, as we had hoped, this didn’t significantly impact the completion rate. The data showed little to no deviation. 😕

Personally, I didn’t feel great about Iteration 3 — it felt more like Dark UX 😬 than a genuine improvement.

But now, one thing was clear: we needed to make UX changes in the second step of the filter flow (see image 1.2). The problem? We had no idea what to change or how to approach it. 🤔

So, I created iterations to simplify the UX of the second step as well — designs that we could implement and test quickly. For example, I made the filter page vertically scrollable, displaying all options under their respective parameters. However, the results were disappointing. No significant improvements, no breakthroughs.

It felt like we had hit a dead end. 😔

🔥 Yes Yes Yes

At this point, I was going through a designer’s block 🧠, and it was clear that without user research, no new ideas were going to emerge. We had already implemented and tested everything we could think of quickly, and now we understood exactly what kind of questions needed to be asked.

Me and Chirag (PM) Requested permission from stakeholders to conduct at least 2 days of user research 🙏. The argument was simple: we had exhausted all small experiments, and big experiments without research would be too risky.

Convincing stakeholders took some effort, but we explained how blindly jumping into big experiments would waste a lot of time ⏳ on design and implementation, while just 2–3 days of user research could save us weeks.

Permission granted! ✅ And off we went to Chomu, Rajasthan. 🏜️✨

🔥 Usability testing

I called one of my power user, Girish Ji, to explain the purpose of our research. He kindly arranged meetings with people who could provide valuable insights. 🙌

We wanted to understand what people do the most when using smartphones. Some participants were given specific tasks, like finding a “4-liter and fifth-lactation Murrah buffalo” 🐃 (since buffalo listings were showing beyond 2 folds). Others were given complete freedom to browse and search for their preferred livestock.

This testing revealed several insights that could never have been discovered while sitting in a chair. Among all the findings, three key points gave a new direction to our experiments:

Blind Spot:
Our UX copy and filter button highlighting experiments failed because users had a blind spot for the upper part of the screen. I noticed that users’ eye movements were immediately drawn to the first listing pictures, compelling them to start scrolling through the listings. 📸

However, this exploration was typically short-lived, as users only scrolled through 2–3 listings before discontinuing. 🛑

This lack of visibility likely explained their limited engagement with the filtering system, as their attention was captured solely by the listings below.

The Search Method:
During user interviews, I observed that users were hesitant to use the old filter due to its tendency to produce surprising and unanticipated results. For dairy farmers, buying animals is a critical and meticulous process, similar to how urban individuals purchase cars 🚗.

They have specific criteria and need the ability to create multiple combinations of parameters to find the most suitable animals.

However, the old filter system was:

Time-Consuming: Changing even a single parameter required users to click on filter button on home page>on filter page click on left side>then click on right side>select options>click on show results, making the process feel redundant and frustrating. ⏳

Non-Interactive: For instance, if a buyer used three parameters and received no results, there was no indication of which parameter was causing the issue. This lack of interactivity left users confused and discouraged. 🤷‍♂️

Category Confusion:
Users couldn’t figure out how many categories of livestock 🐄🐂 were available on the app. This lack of clarity made navigation and discovery more challenging. 🤷

These findings gave us the much-needed clarity and direction for our next steps! 🚀

💡 Inception

The first improvement was crystal clear:
We needed to display categories upfront on the homepage so that users could easily understand the variety of animals available. 🐃🐂🐏

This change aimed to:
✅ Provide instant clarity about available options.
✅ Reduce user confusion by offering a structured overview of animal types. 🚀
✅ Create a more intuitive starting point for buyers to explore listings efficiently. 🔍

Now, the goal was to design a filter where buyers could effortlessly apply and change parameter combinations repeatedly to find livestock without any hassle.

At that time, most filters in the market worked in the same old way — select parameters, hit “Show Results,” and then go back to make changes. It was tedious and time-consuming.

That’s when the idea struck: LIVE FILTERS. 🚀

The concept was simple yet powerful:

1. Instant Updates: As soon as users clicked on any parameter, the listings would update live — no need to go back and forth, applying parameters and hitting “Show Results” repeatedly. 💨

This approach aimed to make the search experience faster, smoother, and frustration-free. ✨

🎨 Live Filter Ideation

I began working on some potential wireframes to bring the Live Filter concept to life.

Gradually, a few potential designs started taking shape, paving the way for a refined and user-friendly filter experience.

🚀 Design, Dev & Deploy

To decide how categories should be displayed on the homepage, two final designs were prepared and launched for A/B testing. 🖼️🎯

Design B outperformed Design A significantly 📈 and even coming quite close to the earlier experiment’s success (check image iteration 3 above for reference). 🎯

Design A didn’t give significant results for two key reasons:

1. The blind spot issue persisted, as users still overlooked the top portion of the screen.
2. It became clear that our users were more interested in visual cues than plain text — they preferred to see rather than read. 👀

Seeing the potential in Design 2, we decided to expand its release gradually, testing it in more user buckets to refine and validate its impact.

The journey didn’t end here. The most difficult part was yet to come:

1️⃣ Designing the Live Filter: It needed to be intuitive enough to show users instant results as they adjusted parameters. ✨

2️⃣ Real-Time Feedback: The filter had to produce results within milliseconds, ensuring there was no noticeable delay to maintain a smooth user experience. ⚡

3️⃣ Technical Feasibility: This introduced significant technical challenges, requiring seamless backend and frontend integration to handle rapid data processing and updates efficiently. 🖥️

While designing, my primary thought was to avoid creating something overly complex. The simpler the design, the lesser the learning curve for the users. 📚 Moreover, simpler designs also allow quicker experiments and faster results.

Here’s how the new flow worked:

• As soon as the user clicked on a category (e.g., “buffalo” 🐄), a bottom sheet opened, displaying all the listings for that category.
• If the user wanted to refine the listings further, they could open any parameter (e.g., “Age,” “Milk” “Lactation”).
• Clicking on an option instantly updated the results in real-time, eliminating unnecessary navigation. ⚡

📖 Design explanation

Flow 2: Placing the Filter Options
In the Filter Page (Flow 2)☝️, I deliberately placed the options below rather than at the top in a dropdown format like this 👇

Reason:
The homepage blind spot issue (users overlooking elements at the top) was a critical learning. And I kept the options as close to the user as possible, ensuring they faced no issues while interacting, and felt encouraged to apply filters 🎯.

Flow 3: Navigating the Filter Options
This stage proved to be the most challenging because it required seamlessly integrating filter options while maintaining a clear visual connection to the listings.

Initial Problem:

1️⃣ In Flow 2’s detailed view, opening filter sub-options would overlay the single visible listing, obstructing it.
2️⃣ If a buyer selected a sub option, only one listing appeared to change, which didn’t clearly convey the impact of their selection.
3️⃣ A listing counter in the top-right corner was added to show changes numerically, but it felt detached and non-interactive.

Dynamic Solution :

💡 Grid View Transition:

• When a buyer clicks on an option to open sub-options, the listing layout automatically transitions to a grid view, showing more listings in a compact format.
• This ensures buyers can visually track the changes across multiple listings as they adjust parameters.

💡 Seamless Back to Detailed View:

• As soon as the buyer clicks cross, the grid view reverts to the detailed view, resuming the natural browsing experience.

💡 Interactive Listing Selection:

• In the grid view, if a buyer clicks on a specific listing, the interface transitions smoothly into detailed view and
• The screen auto-scrolls to the specific listing, allowing buyers to stay oriented and explore further details intuitively.

Benefits:

✅ Enhanced Clarity: Buyers can now see how their selected sub-options influence multiple listings simultaneously.
✅ Visual Engagement: The interactive design bridges the gap between filter adjustments and results exploration, improving user confidence.

This dynamic flow creates a seamless blend of interaction and usability, ensuring buyers feel in control while navigating their choices. ✨

Why a Bottom Sheet Instead of a New Page?
Our users often clicked things accidentally or entered flows without fully understanding them. This led to panic, with users repeatedly pressing the phone’s back button, sometimes even killing the app in frustration.

By using a bottom sheet:
1️⃣ Users stayed in the same context, reducing confusion.
2️⃣ It was easier to exit or modify parameters without the fear of “getting lost.”
3️⃣ This allowed a quick preview without overwhelming the user.

This approach balanced usability and simplicity, tailored specifically to our users’ behavior.

If the user began scrolling through the listings, the bottom sheet transformed into a full-page view.

This behavior served two purposes:
1️⃣ User Intent Confirmation: Scrolling indicated that the user was consciously engaging with the content, not accidentally landing there.
2️⃣ Enhanced Usability: Converting the bottom sheet into a full page provided a better viewing experience for users committed to exploring the listings.

By seamlessly merging the bottom sheet and full-page view, we ensured that users could transition naturally between exploratory and intentional browsing without friction. ✨

On the technical front, the team worked diligently to ensure that the results updated as quickly as possible. However, a slight delay was inevitable. Instead of letting this delay go unnoticed or covering it with a generic buffering animation, we implemented a smart microinteraction to bridge the gap. ⚡

• Purpose of Microinteraction:
  The microinteraction provided visual feedback to users, clearly indicating that their actions (like selecting parameters) were actively changing the results.
• Impact:
  1️⃣ Users felt confident that their inputs were being processed.
  2️⃣ It created a seamless transition between user actions and updated results, making the delay feel almost nonexistent.

🚀 The Final Push

After achieving a significant jump in filter usage from 12.3% to 23% and improving our CWA metric from 16.5% to 24.8%, I was thrilled.

However, I couldn’t shake off the feeling that the homepage blind spot might still persist. If resolved, it could further boost filter clicks, filter implementation, and ultimately, CWA.

But the big question was:
👉 How do we guide users to notice the categories at the top without using conventional methods like tooltips that our users tend to ignore?

The Breakthrough Moment

While tweaking the bottom sheet’s black background opacity for focus, an idea struck:
Could we control the user’s eye movement right as they land on the homepage?

The Solution

Instead of relying on old, ignored methods, I designed an engaging visual cue:

1️⃣ Focus Mode:

• Upon landing on the homepage, users would first see only the top category section.
• The rest of the screen (listings) would be covered with a semi-transparent black overlay, drawing attention upward.

2️⃣ Full Page Reveal:

• After 400ms, the black overlay smoothly disappears, unveiling the entire page.
• This subtle animation naturally guides the user’s eyes to the categories first, ensuring they don’t miss them.

🚀 Massive Success

This innovative approach to tackling the homepage blind spot led to remarkable improvements:
✅ Filter Click Rate skyrocketed to 63%!
✅ Filter Implementation Rate jumped to 54.6%!
✅ Success Metric (CWA) — the rate at which buyers called sellers — hit an all-time high of 39.3%!

These results were beyond what any of us had imagined. 🎉

A huge shout-out to the amazing team behind this journey: Chirag Gupta, Akshat Bhargav, and Ayush Arora — your collaboration and dedication made this milestone possible. 🙌

Special Thanks to Anurag Bisoi 🙌