Rooulette - Case Study

Roo-ulette

A personal project to solve indecisiveness when choosing what to eat, Roo-ulette is a pick-your-meals feature for when everyone feels like eating ‘anything’.
‍
Disclaimer: This project is not affiliated with Deliveroo in any way. This is only a personal project, intending to use Deliveroo as a grounded example.

Role

Product Designer

Timeline

2 Weeks, Solo Project

Project Stack

Figma, Adobe Creative Suite

The Problem

Many others (like myself) could never decide what to eat.

With the regular occurrence of covid lockdowns, deliveries have become my main option of getting my meals. However, being an indecisive and picky eater has turned this simple action of deciding what to order into a painful experience of endless scrolling, browsing and deciding.
‍
A typical conversation between me and my housemates almost every dinner time

My high level goal was to...

Help mitigate the decision making aspect from the user’s flow.

I used the word mitigate instead of remove as I wanted to ensure that the user still possess some level of control over their final choice of order. I initially thought of completely letting the feature do all the deciding, but later on discovered that the ideal functionality needs to allow users to define constraints to the selection first.

Rooulette

Roo-ulette now makes it ever easier for you to decide what to order, with even less steps.

Available when you need it

Roo-ulette shows up as a friendly prompt that triggers after a long time of endless scrolling.

Favourites and My Lists

You can now save restaurants into your own ‘Lists’ which could later be used for your roll. You can save a restaurant into multiple lists as well.

Rolling with Filters

Customise your roll using Deliveroo’s extensive filters to fine-tune the restaurants being included in your roll. Watch your roll gets updated as you add more filters.

Rolling with My Lists

Not feeling adventurous? You can instead choose to spin the roll from your own custom lists of favourite restaurants that have proven themselves time and again.

Roll away, Order Away

To nudge you into making up your mind, Roo-ulette is only available to spin for a limited amount of times and gets recharged over time.

Take it for a spin yourself

Tonight, Roo-ulette says I should be getting...
‍
🍟🍕🌮🥘🍝

How I did it👇

Outlining a research plan

Although this project idea came about from a personal pet peeve, I wanted to to see if anyone else shares the same struggle, but has different needs and hence requires different set of solutions. With this brief in mind, I kicked off the project with user research —

Contextual inquiry followed by user interviews.
Desk researches on topics related to collaborative decision making and several articles on Hick’s law.
Short survey sent to a mutual foodie group for quick quantitative data.

Competitive Analysis

Scoping out the players

Interestingly enough, there are plenty of people out there pulling their hair out over this issue as well. And what’s even more interesting is that the struggles and needs were also quite consistent across the users.
‍
Through a quick desk research, I've discovered abundant of these tools that are attempting to solve this problem. After several sessions of competitive analysis and testing, I discovered that almost all were lacking the 3 key functions that users deem most important according to our research insights.

What I found from this...

Key Insight 1

Remove the decision making from my hands
‍Most of the examples I’ve tested would only help you decide up to the type of cuisine or meal category (Thai, kebabs, etc.), but not down to the restaurants themselves.

“Sure it tells me to get Thai, but from where? That’s another level of deciding”

Key Insight 2

All the information to help us decide �
‍A majority of users have a hard time deciding because they require a lot of additional affirmation before committing to a decision. In this context, the ‘affirmation’ is in the form of any information that helps build credibility to the suggested restaurants:

Are these restaurants well-known? Highly rated? Established?
Are the food going to be good? Portion worth the money?
Do they even do delivery?

“Should we really go with this? I’ve never even heard or seen them before”

Key Insight 3

Different culprits for different occasions �
‍Most users already have their own set of go-to restaurants depending on the occasion or mood, and this list rarely changes. So customisation is highly important to help cover the times when the filters are not specific enough, or when users want to decide just within their comfort options.

“It’s Friday night, time to bring out the guilty pleasures”

Solving the Problem

Identifying existing structure

As an add-on to an existing platform, the first step is to understand how Deliveroo’s app currently works. To summarise what I did at this point-

Investigated the current user journey of completing an order. This helps inform where to best introduce the new user flow and minimise the additional steps required.
Analysed the app’s current feature sets of supporting browsing, discovering and eventually decision-making. Deliveroo already offers good categorisation and filtering so we’ll work together with that feature.
Design and task analysis to keep the functionality and branding consistent for a seamless and familiar experience. It should feel like a complete experience and uniquely Deliveroo.

Prototyping

Jumping to Mid-fi Prototype

Training myself to work around the given time limit, I made the call to skip paper prototype and jumped straight to digital. This was done also due to the consideration that screen interaction is a key component to communicate the feature’s experience and flow. Therefore, a higher level of fidelity is needed to get more reliable testing results.

Feedback, Usability Tests and feedback

First testing session

Early testing results were promising 🙌
Users were able to navigate across screens and functions with little guidance, suggesting that the content hierarchy and user flow were well intuitive.

However, many noted that the experience felt “incomplete” as the cards appear to be the same, as if nothing was ‘decided’ for them. To uncover more testing insights, I quickly moved on to high fidelity, adding images and the rest of the visual elements.

95% of users were able to complete the given scenarios or ordering with minimum required steps and mistakes.
Greatly improved the conversion rate of placing an order by reducing the amount of decision making time by 70–80%
User flows were consistent and seamless. No hair pulling done!

Reflections

Check your bias - I took on this very personal project to challenge myself on controlling bias within my design. And although a lot of my personal needs aligned with the users’, I managed to only rely on research-backed insights when making the design decisions, rather than my personal preferences.

Process is just a guide - For my other projects, I’ve relied rather heavily on industry processes like Double Diamond and found myself sometimes doing things for the sake of following the process. However this time, I tried following where my curiosity would take me and only do things that would get me what I needed to know. As a result, the outcome felt more rigorous and extensive than I even expected.