Donut Shop Ordering App
A UX case study conducted in partnership with a local food business, designing and validating an MVP mobile ordering experience to reduce decision paralysis in highly customizable ordering through choice architecture, behavioral design, and streamlined user flows.
At a Glance
- Identified decision paralysis and friction points through in-context behavioral observation
- Redesigned the ordering flow using behavioral economics and UX heuristics to reduce cognitive load
- Built a functional MVP mobile ordering prototype for rapid testing and iteration
- Demonstrated how improved UX can reduce cognitive load while preserving choice and business flexibility
Background
My favorite local donut shop offered a “build your own donut” bar that customers loved and struggled with. A massive chalkboard listed dozens of icings, toppings, drizzles, and add-ons in a single undifferentiated block.
The abundance felt generous, but in practice it produced confusion, hesitation, and regret. Kids froze. Adults overthought. Group orders felt stressful. Customers routinely asked clarifying questions or changed their minds mid-order. From a product perspective, the chalkboard menu functioned as an unstructured interface, forcing customers to perform unnecessary cognitive work at the point of purchase.
The Problem
- Too many options presented at once, with no structure
- Unclear rules (e.g., “choose a topping” — one or many?)
- No way to visualize the final donut before ordering
- High decision fatigue for kids, groups, and neurodivergent customers
This environment mirrored classic findings in behavioral economics: when choice sets become too large and poorly organized, people stall, default poorly, or walk away dissatisfied. These issues created clear UX failure modes: stalled decisions, reliance on staff intervention, order revisions, and post-purchase regret—all of which impact throughput, satisfaction, and perceived quality.
Design Framework
I approached the redesign using principles from choice architecture, cognitive load theory, and applied decision science, treating the ordering flow as a decision system rather than a static menu.
- Thaler & Sunstein’s work on nudges and defaults
- The Jam Study (Iyengar & Lepper) on choice overload
- Miller’s 7±2 limits on working memory
You don’t have too many choices. You have too little structure.
The Solution
1. Start With Delight, Not Decisions
The home screen opens with a playful animated donut rather than an immediate decision. This moment of delight slows users down and reduces perceived pressure.
2. Structured Customization
This step functioned as feature-level constraint design, guiding users through a sequence of low-effort micro-decisions instead of a single high-cost choice. Toppings were grouped into seven clear categories (icing, sprinkles, drizzles, etc.). Users could choose one—or none—from each group, dramatically reducing cognitive load.
3. Real-Time Visualization
Real-time rendering acted as immediate feedback, reducing uncertainty and improving decision confidence—key drivers of conversion in customizable products. As users selected options, the donut updated live using layered PNG assets. This eliminated ambiguity and reduced regret by making outcomes visible.
4. Escape Hatches for Decision Fatigue
For moments of overload, I added:
- A playful Randomizer that generates a donut automatically
- A gentle timed nudge toward curated “House Specials”
These elements served as intentional fallback mechanisms, allowing users to progress even under high cognitive load without abandoning the flow.
Checkout & Completion
The cart, checkout, and receipt flows were intentionally simple, providing clear feedback at every step. Subtle animations (button shakes, confetti) reinforced progress without adding friction.
Outcome & Reflection
Informal usability testing and in-situ observation suggested reduced hesitation, fewer clarification questions, and higher confidence at checkout compared to the chalkboard baseline. Users (especially children and neurodivergent users) responded strongly to the Randomizer and real-time visual build, often expressing excitement and confidence in their final choices. In contrast to the chalkboard menu, the structured flow reduced hesitation and second-guessing. Customers spent less time deliberating, asked fewer questions, and appeared more satisfied with what they ordered.
Although the project did not progress to full deployment, the MVP validated key assumptions about choice structuring, visualization, and default design under real-world constraints. It revealed how choice architecture, visualization, and defaults can meaningfully improve decision quality without limiting freedom.
Working in a small, informal team also shaped the design process. The problem space was initially vague (“the menu is overwhelming”) and had to be translated into concrete, testable design decisions under tight time and technical constraints. This required prioritizing interventions with the highest behavioral leverage, building only what could realistically be prototyped, and making tradeoffs that balanced user benefit with business feasibility.
The experience reinforced a core product lesson: good design is rarely about ideal solutions. It’s about making principled choices within real constraints, and knowing which details matter enough to build.
Key Takeaways
- More choices aren’t better. Better structure is
- Defaults and visualization reduce regret
- Good UX protects users from fatigue, not from fun
- Behavioral UX interventions can be low-cost, high-leverage product improvements, especially in resource-constrained environments
Explore the fully interactive prototype here
