No idea, no problem!

A lot of times we are hungry, but stuck in the indecisiveness of not knowing what to eat, or simply bored of what we have had everyday. Just let this fun LINE MAN Mystery Meal feature choose for you so you can think about other things instead!

---

Scope of the project

Our goal is to create a new feature for LINE MAN Food that assists users who have trouble deciding on what to eat. This will include adding elements in related screens as well as modifying backend functions, but we're mainly focused on designing a user-friendly interface.

Project statusAccepted 

The project is submitted to LINE MAN Wongnai in 2022 as a candidate assignment for LMWN Junior 2022 Programme and has been presented in the candidate interview.

Before we begin...

First, we need to know our users in order to produce suitable solution.

1 Data from THUMBSUP Media | 2 Data from Forbes and Fountainhead |3 Data from Post Today | 4 Data from Business News Daily and Emerald

Empathise & define

Here is the thing: normally, users have their own go-to favourite restaurants, and they have what they want to eat in mind. That’s because the user already tried the meal and know it will be as good as always.

However, the problem arises when our user want to try something new from different places.

"There will be a lot of dishes to try - but which one should I try this time?"
"How can I know if this feature is wonderful or just a gimmick?"

Design objectives

So, what we need to do in this new feature is to...
"Provide users new menu to eat with most satisfactory experience and least effort from them."

and here are what we will need in our new feature:

Let's wind back a bit

Earlier, we have identified two complications: "what should our users try" and "how to make the choice most satisfiable"

Let's begin with the first one.

Randomisation is the key

How do we make sure our users get the best meal recommendations? Simply suggesting what others have already ordered isn't exciting or helpful. We need a more creative solution - a randomization feature that will suggest what they should eat. Of course, we'll need to use user data to make sure the suggestions satisfy our users.

But let's focus on making the presentation of the randomized meals exciting for now.

Randomisation methods

It is worth noting that before I ended up with this idea, I have explored other randomisation options as well.

Securing user satisfaction

Great! We now have a solution to our problem "What menu should our users try": Users will be provided a mystery meal feature (similar to the mystery box), which will pick a randomised menu for them.

However, here’s another problem we need to solve: How can we satisfy them by our choice?

Good & bad selections

Left: "So lots of customers tipped the cat and the 'menu' is now in #3 popularity", "Snacks tip for Mocha the cat"
Right: Recommended menu - "Tip for Miso the cat", "Tip for Black Jelly"

You might have seen these already...

Although I'm a big fan of cats and dogs, the randomisation algorithm might end up choosing pet-related items as one of the menu choices, which defeats the purpose of the feature that is meant to provide users with food choices.

Furthermore, some restaurants serve both food and drinks, and it's possible that some dishes might not be up to par. Using a simple "pick one" algorithm might result in serving subpar dishes, which could ruin the user experience.

We need a more sophisticated randomisation algorithm to ensure that our users are served the best.

Preferred randomisation

In our randomisation algorithm, we will need it to be somewhat preferable (or biased) in order to provide the most favourable options for users. This can be achieved by using information from LINE MAN database as well as user-selected preferences as inputs for our algorithm.

Building food database

Users will receive a notification to rate their meal and ordering experience on a 1-5 star scale 30 minutes after completing their order. If they accidentally close the rating screen or finish their meal early, they can access the interface through their completed order history.The meal rating creates a rating database for each menu and provides more accurate data than order amount. When the database is robust enough, the star rating will be used to incorporate menu selection. The ordering experience rating allows users to provide feedback on new features and the application itself

Merge into existing UI

Icon design

However, these designs won’t work for several reasons. Mainly, as our feature will autonomously do something for our user, question-mark-based icons will mislead them, thinking it’s a help or food guide. Pairing those with box illustration did not work as well since it’s either too vague to be understood or too detailed to be noticed.

The icons are now a little further from the idea of mystery meal, having no box and no indecisiveness illustration. But on the brighter side, the new icon won’t mislead users into thinking the feature is a guide. The issue here is even the illustration is simplified, the icon still looks cluttered and noisy, and some even leads to another different meaning.

These icons are meant to convey the idea of "tap this and see the magic", which is close to what we intended: "If you press this button, LINE MAN will (magically) choose what I will eat."

The icon uses illustration from FontAwesome 5 assets.