Giftbox app

The goal of this project is to design a giftbox app, similar to the French apps Smartbox and Wonderbox, but with a redesigned user experience. This is a work in progress.

Specific problem to solve

Since gift box apps provide a wide variety of activity choices, users find it hard to select what they will do with the box. The existing filtering system that asks users to choose categories is too rigid to match the exact tastes of the users.

Proposed solution

It is possible to improve the user's experience for the selection process by asking them to answer a sequence of questions. Based on the answers they give, it is possible to narrow down the selection of activities.

The algorithm used to ask the questions could be similar to what the Akinator app uses. The goal for the algorithm is to roughly divide by half the number of activities that are left to choose from each time the user answers a question.

How is it different from filters?

While basic filters already exists in similar apps, the proposed system is better in the following ways:

• The selection process is interactive. In particular, the users do not have to choose themselves what to filter on, the algorithm can choose the best questions depending on the answers already given.
• Users can answer in an ambiguous ways (eg. not sure, quite okay): it makes it easier to rank the options by giving points to the activities.
• It is possible to ask questions that would not match any existing filters. For example, the question "Which one would you prefer among these two places?", would not only restrict the selection to activities in the selected place but would give bonus points to activities in similar places.
• The algorithm can be improved over time by using machine learning techniques. For example, it could learn which questions were the most relevant to select the activity that the users chose in the end and improve the sequence of questions. It could also make connections between types of activities that are different but enjoyed by the same type of people, and therefore suggest alternative activities even if they do not match the answers exactly.

Possible monetization

Some companies could pay to have their activities featured in priority in the app. A limited number of activities could get bonus points in the algorithm, which would make them more likely to be selected (provided that they match the user's preferences to some extent). A "featured" tag could be added to these activities for more transparency.

Companies could also make limited time offers to get their activities featured in a specific "special offers" section of the app.