An agent-based system for modeling users? acquisition and retention in startup apps

Startup companies boost the quality of everyday life in almost all dimensions, and their products and services are of relevance everywhere. One of the most important goals that startups pursue is to increase the number of their users quickly. Users are of two types, new and returning. The present study presents an agent-based model to simultaneously deal with these two types of users by placing them in a preferential attachment network to interact. In this model, new users can be added at each time step according to word-of-mouth (WOM) and marketing activities. To define the retention probability for an agent, a set of real users in the records with the same properties as the agent are looked for to check what they have done in the same situation. To validate and test the model, the agent-based system is first thoroughly verified and then applied to the real data of a startup in the game app industry. After the experiments on a real scenario, the best decisions are made about the users to focus on or incentivize, and the best combination of acquisition and retention policies is adopted. The results show that user retention on the early days of adoption is better than the acquisition of new users. In this regard, acquisition should be focused on when the retention is in an acceptable state. Furthermore, the highest increase in the number of users occurs when there is a good balance between acquisition and retention.

Automated summary

Startup companies aim to quickly increase their user base, which consists of new and returning users. The study proposes an agent-based model to handle both types of users simultaneously. The experiments show that focusing on user retention in the early days is more effective than acquiring new users, and achieving a balance between acquisition and retention leads to the highest increase in user numbers.