Serviceable Recipe Display

Increasing Order Fulfillment Success Rate for All Day Kitchens
All Day Kitchens (formerly Virtual Kitchen Co) is a distributed restaurant platform that partners with restaurants and food brands to help them make their food more widely available on food delivery services like UberEats, Doordash, and Caviar. In the midst of expanding their kitchens and onboarding new employees, they were experiencing a high rate of order inconsistency.
This article documents the process that I led with VKC including:

Role

Product designer

Duration

June 2020 - Aug 2020

Team

1 designer
1 Biz Ops Lead, Alexis

Responsibilities

Experience design
Design Research
Project Management
Visual Design
UI/UX Design
Prototyping & Testing

Context

I was tasked to create an order fulfillment system that would help increase order success rate and improve QA by visualizing recipes and training material on a physical display and building upon employee onboarding. I worked alongside Alexis, who was the biz ops lead accountable for the project.

The Problem

VKC partners with over 15 restaurants offering over 60 menu items to create “Food Hall” on Uber Eats, Doordash, and Caviar. Food Hall is like any other restaurant on food delivery apps except you can order a Mac and Cheese from Homeroom AND a Poki Bowl from Poki Time all in the same order! Unfortunately, this means that order fulfillment can get messy, especially since there are so many recipes and restaurants to keep track of. Customers were complaining that their orders were often missing sides and add ons and the presentation of the meals were not consistent. Our goals for this project were to create a serviceable dashboard that would accomplish the following: 

Local Food Hall(VKC) in the UberEats mobile app

Research

To stay within budget and scope we implemented a lean UX approach. In the initial round, I conducted preliminary research on the onboarding material and organized the requirements into a set of concept prototypes for user testing. In the second round, we took the insights and feedback and iterated upon a final prototype.

Research Questions

With this and the aforementioned goals in mind, I sought to understand the following research questions: 

Preliminary Research

VKC’s onboarding material was fairly simple. For each restaurant there is a slide deck that introduces the restaurant, the dishes and sides available, and specific instructions on how to prepare each dish. Accompanying the deck is a video that walks employees through how each dish is made step-by-step. Since all of the dishes are pre-made, all meals are prepared via a large steamer oven or a turbochef (an industrial convection oven). Since there is already a plethora of information available to the kitchen staff, my task was to figure out how to simplify and display the information in the kitchen for both experienced and new staff members.

VKC's onboarding material

The major steps of preparing any given dish at VKC can be broken down into 4 steps: Item Gathering, Preparation, Assembly and Bagging. For this project, we decided to focus on Preparation and Assembly. After studying the onboarding material, I came up with a set of initial concepts for a dashboard to be used as that considered all aspects of food preparation such as:

the onboarding materials were the building blocks of our prototypes

Our many explorations!

User Interviews

I interviewed 3 of VKC's current kitchen staff from varying kitchen locations. Interviewees encompassed a wide range of experience from new trainees to kitchen leads and testing was conducted on-site at VKC's kitchens. I asked some contextual questions and conducted 2 exercises.
In the first exercise I asked participants to perform a mock training with me. I followed up with a second exercise and asked them to craft their own recipe instructions. Lastly, I showed interview participants a set of 3 prototypes to help them conceptualize what a recipe and order management system could look like while I collected their impressions.

Interview Participants
Maria,
staff trainer, employed 1/2 year
Adaly,
kitchen staff, employed ~1 year
Michelle,
staff trainee, employed 2 months

Read Full User Research Summary

User Persona

We crafted a user persona from the insights gathered in the interviews. While we considered both new trainees and experienced kitchen staff in the design process, we decided to optimize for the experienced employee for 3 reasons:

Our end user

Exercise 1: "Train me" Bodystorming

To help kitchen staff get in the mindset of instructing, I asked participants to pretend as though they were training me, a new staff member, through the making of an order. In the process of teaching me how to make a recipe in real-time, interviewees were able to highlight which steps were most important, include crucial details that might not have been on the training guides, and provide guiding best practices along the way.

Exercise 2: Craft Your Own Recipes

Interviewees were then instructed to create their own recipe dashboards on a provided template. They were encouraged to create as many steps as they felt were necessary.
To view the full artifacts, see the full interview summary

Prototype Feedback

One of the biggest unknowns that we discussed was how much information to display. Below are the 3 prototypes that showed varying levels of detail and media type along with the overall feedback collected.

Cherry-picked action items included 1-2 gifs or short clips that displayed the most important steps for a dish

Ingredient-based key steps outlined all major steps in the process of preparing a given dish

Key steps included 4-5 images that outlined the most important steps for a dish

Illustration
👍🏼
easier to build a library of content that is reusable as catalogue grows
👍🏼
easier to maintain visual consistency
needs to be crated from scratch
Requires more resourcing (illustrator)
GIFs / Screencaps
👍🏼
Helps with memory recall
👍🏼
can be built from existing onboarding material
👍🏼
simple to create and doesn't require extra resourcing (converting video clips to GIFs) 
harder to maintain visual consistency as recipe and restaurant catalogue expands

testing media type

Key Actionable Insights

 We analyzed and distilled the results from our interviews into 5 key actionable insights. With a better understanding of the training process and how kitchen staff operate in the kitchen, we utilized the insights to ideate and refine upon the final prototype.

Designing the Prototype

Using the "Key Steps" prototype as a skeleton, Alexis and I put together a feature roadmap that helped us prioritize all the features necessary for a V1.

Wireframes

I sketched out a number of concepts through low-fidelity wireframes and exploring a variety of different ticketing interfaces. I validated my designs with a group of designers from my network and the consensus was a version with a side menu that delineated the incoming tickets from the recipe instructions itself. This allows for easy ticket navigation and encourages users to complete tickets in order from most urgent to least. Order completion CTAs are found on the side menus in consideration of rush hour when kitchen staff are sequencing and multiple tickets may need to be ticked off at once.

Final prototype & handoff

I presented the final prototype to Alexis and the ops team at VKC for a final handoff. The prototype was a simple but clear interface that allowed for easy navigation through tickets and recipes.