June 2020. The world was locked down. Local businesses were struggling to reach customers. Delivery platforms existed for restaurants and large retailers, but the small neighborhood vendor – the grocer, the salon owner, the local restaurant – had no digital presence and no easy way to get one.

This wasn't a hypothetical problem statement pulled from a strategy workshop. It was the daily reality for millions of small business owners, particularly in India and across developing markets. Customers working from home wanted to order from their trusted local shops. Vendors wanted to keep selling. The bridge between them didn't exist.

I joined this project as the design lead with a mandate that was broader than the title suggested. The team was small – five people across four disciplines – and in a team that size, boundaries between roles dissolve quickly. I was responsible for the entire design surface: product strategy, UX research, interaction design, visual design, UX writing, and co-leading marketing.

The team:

• Design, Research, UX Writing: Me
• Marketing: Pranav Patil and me
• PM: Nikhil Patil
• Engineering: Separate team handling Android development

What was at stake: This was a real product for a real market – not a concept project or a portfolio exercise. If the app didn't work for vendors (who had zero technical sophistication and deep scepticism about online platforms), it would fail regardless of how polished the customer side looked. Both sides of the marketplace had to work, and they had to work for people who had never used anything like this before.

The surface-level problem was straightforward: customers need a way to order from local shops, and shops need a way to receive and manage those orders. A two-sided marketplace. The architecture was well-understood – Uber, DoorDash, and dozens of others had proven the model.

But the real problem wasn't the marketplace mechanics. It was the people on both sides.

The Customer Side: Customers were primarily millennials working from home during the pandemic. They needed groceries delivered, but also wanted to book appointments at hair salons, reserve restaurant tables, and order food in advance. These weren't e-commerce shoppers accustomed to cart-and-checkout flows – many were making their first digital transactions with local businesses. The experience needed to feel as natural as walking into the shop and asking for what they needed.

The Vendor Side – The Harder Problem: Local business owners had a fundamentally different relationship with technology than customers did. They managed stock by instinct, not spreadsheets. They communicated with customers through WhatsApp and phone calls, not apps. They preferred cash. And critically – this is the insight that changed everything – they didn't see the value of going online.

It wasn't resistance born from inability. It was a calculated assessment: their businesses survived on relationships, foot traffic, and word of mouth. An app felt like overhead with no clear payoff. The most vulnerable vendors – small, unstable shops – were focused on two things: increasing sales and stabilising their business. Going online wasn't on that list.

This framing changed our entire approach. We weren't designing an app for vendors who wanted to go online. We were designing an app for vendors who needed to be shown – through the product itself – that going online could directly increase their sales and stabilise their business. The value proposition had to be embedded in the experience, not pitched in a marketing deck.

[IMAGE PLACEHOLDER: "The Two-Sided Problem" Diagram]

Two apps, not one.

The customer and vendor experiences were so fundamentally different – different mental models, different tasks, different levels of technical comfort – that we designed them as separate applications sharing a common backend. This wasn't just a UI decision; it was a product strategy decision. A vendor managing orders and tracking stock has nothing in common, interaction-wise, with a customer browsing shops and placing orders. Forcing them into one app would have compromised both experiences.

Research-first, not feature-first.

We resisted the temptation to jump to features. Before a single wireframe was drawn, we needed to understand how vendors and customers actually behaved – not how we assumed they behaved. This meant card sorting exercises, immersion studies, and direct observation. The pandemic made in-person research challenging but not impossible, and the insights we gained were worth every constraint we navigated.

Design for the vendor's mental model, not ours.

The biggest strategic bet was designing the vendor experience around how vendors already think about their business – sales, stock, orders, customers – rather than imposing a traditional e-commerce dashboard. If the app felt like a foreign tool, vendors would abandon it. If it felt like a digital extension of what they already did, they'd adopt it.

Constraints we worked within:

1. Time zone differences across the team (distributed, not co-located)
2. Asynchronous communication as the default – decisions couldn't wait for synchronous meetings
3. Pandemic restrictions on in-person research (limiting but not eliminating fieldwork)
4. Building for users with low technical literacy on the vendor side
5. Limited budget and team size – five people, all wearing multiple hats

Phase 1: Empathize – Card Sorting (Customer Research)

We started research on the customer side in Maharashtra, India. But we didn't run a standard card sort.

Traditional card sorting asks participants to rank cards in order of preference. We modified the method – instead of ranking, we allowed participants to arrange cards as they saw fit, with no imposed hierarchy. This yielded surprising results. We then introduced scenario variations: what if you had more funds? What if you were older? What if delivery times doubled? We watched how priorities shifted under different constraints.

The goal was specific: understand how customers perceived grocery shortages and delivery during the pandemic, and surface needs we wouldn't find through conventional interviews. For vendors, we used the same exercise to uncover their challenges – expired supplies, delivery restrictions, resource constraints.

What we learned:

• Small business owners were primarily focused on two things: increasing sales and stabilising their business
• They perceived little value in going online – this was consistent across vendor types, from street-side stalls to established shops
• There was a widespread cultural perception favouring sales growth over service quality and online presence
• This "blind spot" explained the slow adoption of online business practices among small vendors

This wasn't an insight we could have found in a survey. It came from watching people physically move cards around a table and explain their reasoning. The modified card sort turned a preference exercise into a window on mental models.

[IMAGE PLACEHOLDER: Card Sorting Session Photo]

The actual photo from the case study showing hands arranging coloured sticky notes/cards on a table. This is a real research artifact – use as-is from the original project. Caption: "Modified card sorting in Maharashtra – participants arranged freely rather than ranking, revealing unexpected priority structures."

Phase 2: Empathize – Vendor Immersion

Card sorting told us what vendors said they valued. Immersion showed us what they actually did.

Nikhil, our PM, conducted a full immersion study by shadowing a small business owner in his locality. He embedded himself in the vendor's daily routine: morning stock checks, cash balance verification, customer interactions, order management, delivery scheduling, inventory tracking. He observed the vendor's intuitive stock management practices – not spreadsheets, not even written lists, but mental models built from years of experience. He watched customer interactions, participated in sales conversations, and documented everything in a field diary.

Why immersion over interviews: We chose this method because we needed to see the vendor's operations in their actual environment, not hear about them in a conference room. Vendors describing their workflow would filter it through what they thought we wanted to hear. Watching them work revealed the raw reality: the instinctive stock gauging, the reliance on memory for orders, the preference for cash even when messaging platforms were available.

What we learned:

• Vendors rely heavily on instinct for stock management – they "feel" when supplies are running low rather than checking data. This inspired us to design predictive stock management features using real sales data, giving vendors a digital extension of their existing instinct.
• Vendors were already operating online (via WhatsApp, phone calls) but in an unstructured way. They didn't need to be "brought online" – they needed their existing online behaviour to be formalised and made more efficient.
• Cash payments were deeply culturally embedded. The app needed to support cash-on-delivery as a first-class experience, not an afterthought.
• The importance of shared team experiences – the immersion exercise bonded the team and created a shared empathy that informed every subsequent design decision.

[IMAGE PLACEHOLDER: Paper Wireframes and User Flows]

Phase 3: Define & Ideate

With the research synthesised, we moved into defining the product structure. The two-sided nature of the marketplace meant designing two complete but interconnected experiences:

Customer App Features:

• Browse local shops by category and proximity
• Order groceries and goods for delivery
• Book appointments (hair salons, washing centres)
• Reserve restaurant tables
• Advanced food ordering
• Order tracking and delivery status
• Ratings and reviews for local businesses

Vendor App Features:

• Business profile and service listing management
• Order receipt, management, and fulfilment tracking
• Stock management with data-driven insights
• Sales analytics spanning three months – trend visualisation and stock projections
• Delivery management and scheduling
• Customer communication tools
• Statistics dashboard showing business performance

The vendor analytics feature was a direct response to our research insight. Vendors managed stock by instinct – we gave them data that enhanced that instinct rather than replacing it. Three months of sales trends, stock projections based on actual ordering patterns, visual indicators of what's selling and what's not. We framed it as "your instinct, backed by numbers" – not "stop guessing and use data."

Phase 4: Prototype – Wireframes to High Fidelity

The design moved through classic fidelity stages, but with a twist driven by the distributed team constraint:

Paper wireframes came first – rapid sketching of every screen and flow for both vendor and customer experiences. These were photographed and shared asynchronously for team feedback. The low fidelity was intentional: we wanted to validate flows and information architecture before investing in visual design.

Digital wireframes followed, translating the paper sketches into structured layouts. These were used for early usability validation – does the navigation make sense? Can a vendor find their orders? Can a customer complete a purchase?

High-fidelity prototypes were built in Figma – fully interactive, representing the complete user journey for both apps. The visual design language used a teal/turquoise primary colour palette with warm accent colours, illustrations that felt approachable rather than corporate, and a clean interface that prioritised clarity over decoration.

The vendor app was designed with larger touch targets, simpler navigation, and more visual cues than the customer app – reflecting the different technical comfort levels of the two audiences.

[IMAGE PLACEHOLDER: Hero Image – App Mockups]

The original hero image from the case study showing multiple phone mockups arranged in an isometric grid on a teal background. Visible screens include: the Local Store splash screen with shopping bag illustrations, a Statistics dashboard, a Track Order screen, and various other app screens. The teal colour palette and illustration style are consistent across all screens. Use as-is from the original project. Caption: "Final high-fidelity designs for Local Store – vendor and customer apps."

Phase 5: Testing & Launch

The application underwent usability testing with real users from both sides of the marketplace. The high-fidelity prototypes were tested for:

• Task completion rates (can a customer place an order? can a vendor fulfil it?)
• Navigation comprehension (do users find features where they expect them?)
• Value perception (do vendors see how the analytics help their business?)

The app was built, tested successfully, and launched live – initially targeting markets in Africa where the local commerce model closely mirrored the vendor-customer dynamics we'd designed for.

Quantified:

• Complete product shipped – from research to live application, available for purchase in Africa
• Two full applications designed – vendor-side and customer-side, each with interactive high-fidelity prototypes
• End-to-end design ownership – product strategy, research, UX, UI, writing, co-marketing
• Modified card sorting methodology yielded insights that standard surveys missed – directly shaped the vendor value proposition

Qualitative:

• The vendor analytics feature – born from the immersion insight about instinctive stock management – gave vendors a reason to engage with the app beyond just receiving orders
• The two-app architecture allowed each experience to be optimised for its audience without compromise
• The research-first approach prevented the team from building features that looked right but solved the wrong problem
• The pandemic constraint (limited in-person research) forced creative methodology adaptations that produced richer insights than standard approaches

Personal growth:

• First time leading a product end-to-end as both strategist and designer – learned the difference between designing features and designing a business
• The immersion study fundamentally changed how I think about user research – observation reveals what interviews hide
• Managing a distributed team across time zones with async-first communication became a skill I'd use repeatedly in later roles
• The vendor insight about "not seeing the value" became a principle I apply to every adoption challenge: don't convince users to use your product – make the product prove its own value through use

On designing for two audiences simultaneously: The hardest part of a two-sided marketplace isn't the technical architecture – it's holding two completely different mental models in your head at the same time. A customer thinks in terms of "I want this delivered." A vendor thinks in terms of "how do I sell more today?" The app has to serve both, and the design decisions for one side can't compromise the other. The decision to build two separate apps was the most important strategic call we made – it gave each audience the respect of a purpose-built experience.

On research that changes your assumptions: We went into this project assuming the main barrier for vendors was technical literacy – they couldn't go online. The card sorting and immersion studies showed us the real barrier was perceived value – they didn't see why they should. That's a fundamentally different design challenge. If we'd designed for the wrong barrier, we'd have built a simpler app (solving for literacy). Instead, we built a more valuable app (solving for motivation). The research didn't just inform features – it reframed the entire product strategy.

On working within constraints: Five people. A pandemic. Time zones. Async communication. Limited budget. These constraints didn't limit the quality of the work – they shaped its character. The paper wireframes weren't a compromise; they were the right tool for the speed we needed. The modified card sort wasn't a fallback from in-person interviews; it was a better method that yielded deeper insights. Constraints are design material, not obstacles.

On the pandemic context: This project existed because of COVID-19 – the problem it solved was intensified by lockdowns and movement restrictions. But the underlying need – connecting local businesses with local customers in a digital-first way – outlasts the pandemic. Designing for crisis conditions taught me to build for the urgent need while planning for the lasting one.