We noticed an interesting behavioral pattern on Zomato. Customers would open the app hungry, fall into a 30-minute scrolling maze, look at dozens of options, and then completely close the app. They weren't bouncing because they lacked intent; they were bouncing because they were stuck in an exhausting evaluation dilemma. They wanted to eat, but the sheer volume of choices made deciding feel like a chore.

Fixing this wasn’t a straightforward, linear UX patch. It was a classic convergent research challenge with deeply tangled operational branches—spanning algorithmic relevance, contextual UX, and marketplace liquidity. We couldn't just redesign a filter pill or tweak a single layout; we had to rewrite the underlying recommendation logic, reshape the UI across multiple separate touchpoints, and sync up with multiple teams to incorporate these changes.

Solving decision fatigue meant realizing that "hunger" isn't a singular state—intent is fluid and highly sensitive to external variables. When a customer opens a food delivery app, you are racing against a biological clock. The longer it takes them to resolve their dinner dilemma, the higher their frustration peaks. To help customers navigate their selection hurdles, we mapped the ecosystem broadly and bifurcated our research based on customer intent depth.

High-Intent Customers: Treating Search as an Express Lane

For the one-third of customers who already knew what they wanted, our existing search engine was an obstacle course. They were forced to filter by restaurant brands instead of what they actually wanted to eat. Our goal was to eliminate friction and assist customers in finding exactly what they were craving at that moment.

• Dish-First Indexing: Customers crave dishes, not logos. We shifted the visual hierarchy of search results to index heavily around specific food items rather than just restaurant storefronts. If you searched "Biryani," you saw the best biryani recommendations near you immediately, bypassing the step of clicking into individual restaurant menus.

  
  

• The Relevant Filter: We bucketed critical decision parameters—like "Fastest delivery," "Top Rated," "Trending Nearby," and "Great offers"—and made them more prominent for the customer. No more hunting for boring, tiny filter icons at the top of the screen.
  
  

• Social Proofing as a Decision Trigger: Generic 4.5-star ratings aren't the only factor when you're starving and undecided. We injected dynamic, high-velocity micro-copy underneath listings (e.g., "1.2k people ordered this specific meal here today"). This leveraged social proofing to instantly validate the choice and accelerate checkouts.
  
  

• Changing the Order of Rails on the Homepage: Based on intent depth and the time spent on the app in a given session, we dynamically changed the order of content rails on the homepage to assist customers in their selection. For example, we surfaced the "previously ordered" rail at the very top for customers whose average session duration was less than 5 minutes.

Low-Intent Customers: Designing Contextual Triggers

Then we had the explorers. These were the customers trapped in the 30-minute scrolling loop because a generic list of "Restaurants Near You" presented a massive paradox of choice. For them, we didn't want a system that forced them to weigh pros and cons; we wanted the app to elegantly guide them to a decision.

• Real-World Environmental Triggers: We plugged our homepage algorithm into hyper-local weather and time APIs. If the system detected it was suddenly pouring rain outside, the main hero banners instantly shifted away from standard listings to comfort food pairings—like a highly visual Chai & Pakoda tray. It resolved the "what should I eat?" dilemma by matching their immediate environment.
  
  

• Emotional Nostalgia Loops: We took past order data out of the dry history tabs and pulled it directly into the main feed, framed around emotional contexts (like "Your Rainy Day Favorite" or "Go-to Late Night Spot"). We turned an aimless browsing session into a quick reorder loop.
  
  

• Similar Taste-Buddies: Leveraging historical user preferences, we built taste profiles to suggest specific dishes and restaurants based on what other customers with matching taste profiles were ordering nearby.