
36 hours, in a Hackathon promoted by iFood
Luádyna Almeida, Thiago Lins, Eloísa Schwertner, Pedro Paraguai and Bruna Lopes
This project was awarded Third Place in the competition ranking
Main conclusions
The achievement was the result of an intense 36 hours of work. What we realized in these 36 hours is that we don’t always have to build something extremely complex to serve the user; some things can be simple and at the same time complete. The simple fact of proposing partnerships with influencers could improve the number of users by almost 1 million hits, and this was one of the most praised in the solution.
This was my first participation in an event like this, and what an experience! Being able to learn so much about a company and its goals was very enriching. I meet amazing people, without them, I would never reach the Top 3 in the competition.
The challenge
Hack iFood is offered by iFood, a large foodtech company present in Brazil and Colombia, one of the largest technology companies in the country. In the third edition, they brought a very different theme to the challenge to be made: using artificial intelligence to create innovation for the current platform.
Currently, they are already using artificial intelligence in the restaurant vertical and more. The challenge was to create a feature or product to improve the current system using similar or different technology that they already have. The solution had to be prototyped and coded and had a presentation and demonstration video.
Context
The COVID pandemic has changed the way we shop online; supermarkets register a 57% increase in online sales, through applications and websites. For those who buy online, 37% do so because it is practical, 32% because they do not want or cannot leave the house and 29% choose digital platforms because they offer discount coupons.
In other words, the internet has become the main channel for purchasing food and supplies. Mercado & Consumo, a Brazilian website, shows that 75% of respondents search online for information on diet, recipes, and gastronomy. The most sought-after information is about non-perishable foods, personal care, hygiene products, cleaning products, and perishable foods.
This is because Brazilian homes have adopted their kitchens and cooking as hobbies during the pandemic. According to Youtube, in 2020 in Brazil, 55% of its total hits were on videos on recipe preparation.
At the same time, Brazil sees a huge increase in gastronomy influencers; they are responsible for indicating restaurants and recipes that are taken into account in trips and family moments. Together they accumulate more than 18 million subscribers on Youtube, and many of them run TV shows.
Problem definition
The first step to start defining the problem was to browse the iFood Mercado platform. We did and redid the purchase journey, from different perspectives, to find and map possible problems.
After a quick overview of the problems and a brainstorming session with the whole group, three basic problems were defined in the supermarket store journey. These three issues were considered when thinking about a major purchase, and how we could improve the ingredient selection process.

Survey
To measure the importance of the problems raised, we prepared a small survey containing qualitative and quantitative questions for two groups of users: 1) users who shop for supermarkets through apps and 2) users of physical markets. We chose these two groups because they were able to capture different perspectives on the shopping experience, and in the case of physical supermarket users, understand why apps are not considered for shopping.
The survey lasted two hours, and we accumulated a total of 29 responses. We chose to bring punctual answers but that would bring us interesting feedback, so the time required to generate insights would be shorter. The main insights were:
We could also conclude that users prefer to go directly to the supermarket to buy ingredients for a recipe they have never made, as they do not feel safe buying directly online.
The solution
Seeking to meet the criteria for submitting projects, the solution met the areas of product, technology and business. These three areas should be well articulated to represent an innovation for the company.
Product: A solution that combines the convenience of shopping for ingredients with the desire to discover new recipes. The solution is integrated into iFood Mercado and offers a list of recipes that can be purchased with a single click using advanced technology;
Technology: The product uses artificial intelligence to select the ingredients of a recipe based on the nearest supermarkets, all at once, and show which has the fastest delivery and which has the lowest price for the purchase.
Business: Bring users to the platform by partnering with food influencers (in the solution, we call them “Creators“). The product will showcase selected recipes from YouTube’s biggest influencers, linking the product to their videos
Information architecture
To start building the proposed solution, an information architecture scheme was needed, seeking to understand how exactly the resource would integrate the current vertical of iFood Mercado. It is necessary to say that this solution was thought of as a feature because it was easier to show how it would work; Essentially, the solution is intended to be an API, which can be adopted for any marketplace platform or as a product in itself.
Inside the iFood Mercado, the feature would work:
Prototype
Having established the solution hierarchy, we prototyped our version of the Market page to design the solution flow. In the end, we shortened the specific in-app purchase journey to a maximum of 5 clicks, depending on which path you choose.
To access the prototype, click here.
Product evolution
At the end of the project, we present the best and easiest solution for an MVP. However, the other possibilities for this product show us that it can be improved and evolve much more. Three key points can make MVP better:
1) Suggest recipes based on the products in the market bag. This was the obvious solution, and we believe it could be perfect for an improvement. The AI would read and select the best recipe based on the product and the user could add the rest of the ingredients with one click. Example: You want to buy avocado and tomatoes; the AI suggests guacamole and if you want to put the rest of the ingredients in your bag;
2) Create themed campaigns for each period of the year. In Brazil we have some very typical holidays, so the app will suggest recipes based on that. Example: Turkey for Christmas Eve;
3) Indicate a nearby restaurant that sells the food you want to cook.
Lua Almeida
Made with love, using Framer