DIGICOLLECT
Improving success in task completion by 50%.
The integration of Table Management into DigiPOS's Dine-in view presented an issue. Despite high anticipation for this update, it unexpectedly generated more negative reviews and a decline in user engagement. Servers struggled to navigate the new table management system, which subsequently led to an increase in dropouts.

As the sole designer, I improved the Dine-in view and 'Assign Table' feature by conducting data-driven research, design iterations, and collaborating with development and business teams.
Jump to solution

My Role

UX Designer

Team

1 UX Researcher,
1 Developer

Time

2 Months
(Mar'22- Apr'22)

Tools

Figma, Miro,
Quant Metrics
Match plan with real-world: Easy to assign seats to guests using an integrated floor plan with separated sections and quick actions.
Real-time updates: Enables servers, chefs, and administrators to view real-time updates on restaurant occupancy.
Consistency in language: Easily accessible information about guests and tables, using terms commonly used in the restaurant industry.
Managing reservations, waitlists, and dine-in: Effectively manage guests by integrating an SMS API to handle no-shows.
INTRODUCTION
What is the system & Who are the users?
DigiPOS and Server tasks in Dine-in view
DigiPOS is a cloud-based point-of-sale (POS) system used by restaurants, hotels, and bars. The application offers three different types of user accounts: Admin, Chef, and Server logins. The main functions of the Server login include managing seating, guests, and orders, as well as coordinating with managers and chefs.
Previous version of DigiPOS's Dine-in view screens with Table management
Research
What is the problem?  How does it affect users?
The 'Assign Table' and 'Take Order' functions had the highest number of errors
I collaborated with a UX researcher and obtained information on user behavior from the customer analytics team and identified several issues through heuristic evaluation, including error prevention, system status visibility, user control, freedom, accessibility, and feedback deficiencies. These issues lead to increased errors and incomplete task completion, especially within the 'assign table' and 'take order' functions.
Heuristic Evaluation of Dine-in view
Visual representation of errors from analytics
Activities and output
User Flow Analysis
User Behavior analysis (Fullstory)
Heuristic Evaluation
user & market Research
What did I learn from users and competitors?
Mismatch between system and real-world significantly affected 'Assign Table' task
I leveraged user interviews and contextual inquiries to understand key Servers pain points to complete the 'Assign Table' task. Additionally, servers are facing difficulties in properly completing the 'Take order' task, primarily due to errors in guest and table states in 'Assign Table'.

Incomplete 'Assign table' tasks also affect the collection of data for analytics for Admin login.
1
Inaccurate Table States
The Table status currently doesn't capture 'order' status or the 'guest' status accurately. Current statuses are: "in progress," "occupied," and "available" which users find confusing.
2
Restaurant Plan Configuration Mismatch
There is a configuration mismatch in the restaurant plan within the system. Currently, users rely on paper floor plans to assign seats to guests, which highlights the inefficiency of the system for this task.
3
Lack of Reservation & Waitlist Management
The system lacks the capability to manage reservations and waitlists, thus requiring manual paper entry to manage guests.
Activities and output
User Interview
Contextual Inquiry
Competitive Analysis
PROBLEM STATEMENT
How might we enable servers to complete the task  of ‘assign table’ efficiently and attend to guests at present?
Success Metric: Measuring how quickly the server can assign seats and tracking the number of errors or mistakes made during seat assignments. Target was to reduce the error rate by 5% and increase speed and efficiency by 50%.
Design
How did I improve the existing flow?
Mapping task flow for 'Assign Table'
I began by creating a map of the current server flow, taking into account the system and the pain points identified from user interviews and contextual inquiry. The main issue identified is that over 70% of subtasks in the 'Assign task' process are currently being manually entered on paper, as the system does not support the server's existing flow.
Mapping the current 'Assign table' flow including the tasks which are currently manually entered.
Design principles
1
Mimicking Real-world Familiarity
Users need to feel as if they are interacting with a system that closely resembles the physical layout and processes of the physical restaurant. This familiarity can reduce the learning curve and make the system more accessible for completing the task.
2
Consistency & Clear Visual Representation
Visual cues such as floor plans & table maps should closely resemble the restaurant's layout. This will help user quickly identify table locations and availability.
Early testing & feedback
Through early brainstorming sessions and sketches with the development team, I was able to make design decisions and promote a design thinking culture within the team. Some of these decisions included dividing floor plans based on sections and limiting table shapes to three options (circle, square, and rectangle) for the new version of table management.
Option 1: The unrelated options take up too much space, resulting in a disproportionate amount of space dedicated to relevant information.
Option 2: Arranging relevant information in a proportionate manner, reducing cognitive overload and simplify technical issues.
These collaborative efforts with development team and design decisions made were then incorporated into mid-fidelity prototyping and usability testing with 10 clients.
Customer journey data from Quant Metrics after usability testing
Activities and output
Wireframing
Prototyping
Usability Testing (Quant Metrics)
Collaboration with Development
Based on feedback from usability testing of the Version 1, followed by collaboration with the development team, I created a Version 2 of prototypes. Some key design decisions that improved 'Assign Task' function included:
1
Redefining the first step to selection of guest number
Version 1: Users were confused between adding guests for different buttons, such as reservations and walk-ins.
Version 2: Provided guest number selection as the first step, followed by the type of reservation (reservation, waitlist, or walk-ins)
2
Guest Details & Table Selection function
Version 1: The "Select table" popup created confusion and increased clicks, and only 2 out of 8 participants were able to navigate it correctly.
Version 2: Designed for selecting tables side by side, with clear feedback indicating their availability status.
3
Consistency & Clear Visual Representation
Version 1: The "Select table" popup created confusion and increased clicks, and only 2 out of 8 participants were able to navigate it correctly.
Version 2: Designed for selecting tables side by side, with clear feedback indicating their availability status.
Activities and output
Prototyping
Collaboration with Dev Team
Final Design
How did I solve the problem?
Designing for Familiarity: Dine-in view
1
Matching system plan with physical restaurant plan
By matching system dine-in plan with physical arrangement of seats in the restaurant, the design reduced the learning curve and errors, and make the system more accessible.
2
Selecting seat parallel to guest details
By giving servers user control and freedom to select seats, as well as completing guest details, I improved error prevention and aligned the design with the existing mental model for the server.
Designing for Consistency:
1
Language matching with guests and order states
Making the system language consistent with the restaurant lingo, servers can complete their tasks conveniently and reduce their dependency on external manual entries which increase task time.
learnings
Impact
The data we gathered from early adopters was positive. We witnessed a 50% increase in task completion. The design also increased productivity and received overall positive feedback. The servers expressed that there are fewer errors and they enjoyed using the system. That being said, we are always open to ways to improve.
Designing with limited Time & Resources
The project was my first time taking on the responsibility of improving an existing product's user interface with limited time and resources. Despite the challenges, I found ways to quickly prototype and test key concepts, making it easier for users to navigate and understand.

Other Projects

Marketplace: Solve Student Culture Shock

Persuasive Strategies in E-commerce

An Artificial Intelligence Itinerary for Travel Planning

Instagram Feature