AppRecommender

The research project AppRecommender is aimed at the development of a recommendation and semantic search system dedicated to mobile app stores. This project, sponsored by the Portugal2020 fund with project number 39703, is promoted by Caixa Mágica Software in collaboration with ISCTE-IUL.

Challenge

The proliferation of mobile devices in society has led most businesses to consider the mobile component as essential for close contact with their customers.

In 2017, the Google Play Store had 2.8 million mobile applications available, Apple’s App Store had 2.2 million, and Aptoide currently has over 1 million apps available, creating extremely fierce competition among apps.

In terms of transactions, in 2016, there were 149.3 billion app downloads, a number expected to double by 2020. However, many of these downloads consist of multiple attempts to find the right application, and many downloaded apps are never used. In 77% of cases, apps are not used again 72 hours after installation.

This demonstrates a significant misalignment between the supply of apps from app stores (distribution services) and the demand for them by consumers (discovery). Due to this misalignment and the high competition among apps, Gartner’s predictions indicated that by the end of 2018, less than 0.01% of developers in this market would consider themselves to have achieved commercial success. Furthermore, in the increasingly digital age we live in, 52% of apps are discovered through word of mouth among acquaintances, friends, or family, while only 40% are discovered through searches on app stores.

These inefficiencies make the discovery and distribution of apps a considerable and highly relevant challenge, as they occur in a market that has massively penetrated societies and significantly impact the relationship between companies and consumers.

Solution

Building upon this problem, the AppRecommender project has a strategic objective to research and develop technologies capable of delivering the right app to the right customer at the right moment. To achieve this, it proposes a multi-criteria recommendation system and a semantic search engine.

The multi-criteria app recommendation system will have the ability to populate real-time touchpoints (mobile homepage, web, VR, among others) between the consumer and the mobile app store with applications that align with the consumer’s profile and needs.

On the other hand, the semantic app search engine will aim to understand consumers’ intent during their searches and match this intent, the consumer’s profile, and context with the profile of apps available in the app store, delivering the right apps to the consumer who requested them in real time.

The goal is to optimize current app distribution and discovery services and, consequently, promote closer connections between businesses and their target customers.

Impact

The project aims to have specific impacts on mobile app consumers, the companies promoting apps, and the Aptoide app store through Caixa Mágica Software, the leading promoter of this project, which will bring the results to the market.

For users, the impact will be in terms of increased simplicity, efficiency, and satisfaction in discovering apps due to the optimized alignment between their needs, characteristics, and context with the apps offered by the app store.

For developers or companies promoting mobile apps, the impact will be in terms of improved proximity to target consumers, enabling their acquisition and potential retention, thereby optimizing their commercial success.

For Caixa Mágica Software and Aptoide, the impact will be an increase in the quality of service provided to companies and consumers, as well as an inherent organic increase in apps submitted to the app store, active users, and revenue.