How ByteDance Boosted New User Retention with Incentives and AB Testing

Product Background

Product “M” is a video‑interest recommendation app similar to Douyin, aiming to retain users by satisfying their viewing needs. Its core capability is precise algorithmic recommendation, and the key user action for retention is watching video content.

Incentive System Design to Improve New‑User Retention

To keep new users during the critical first 1‑7 days, the team created a gold‑coin incentive system that rewards users for completing tasks such as watching videos, thereby encouraging more interactions and allowing the algorithm to learn preferences.

In the first‑day activation, users receive cash rewards for completing three one‑time tasks (new‑user red packet, gold‑coin pack, watch video 1 minute) and five ongoing tasks related to video watching. From day 2 to day 7, daily one‑time and three ongoing tasks reward cash and encourage at least nine minutes of video consumption per day.

Thinking Path One

Goal: Ensure new users continue using the product daily and watch more content during the first 1‑7 days.

Explanation: The early experience heavily influences long‑term retention; encouraging more video views increases interaction and improves algorithmic personalization.

Implementation: The incentive system (illustrated above) provides cash rewards for completing specific tasks, with larger rewards on day 7.

Thinking Path Two

Goal: Optimize the incentive system for maximum effect at minimal cost.

Implementation: ByteDance applied AB testing, leveraging its internal data product suite (Volcano Engine) to evaluate different incentive triggers (e.g., short‑video vs. live‑stream contexts) and the optimal amount of gold coins.

Through multiple AB test comparisons, the team selected the most effective incentive strategy, which led to significant improvements in new‑user retention metrics after launch.

How to Identify a Product’s Core Function

Finding the key feature that drives retention involves three steps:

List Important Functions: Product managers enumerate all potential core features.

Locate Key Functions and Behaviors: Import the list into the growth‑analysis tool to see which features correlate with the highest retention; the top‑ranking feature(s) become the core function(s).

Incentivize the Key Behaviors: Use operational or product guidance to encourage users to perform the identified key actions, thereby increasing perceived value and long‑term retention.

Conclusion

The case demonstrates how, in the 0‑1 stage, ByteDance employed a structured incentive system and rigorous AB testing to drive new‑user growth for product “M”. The growth tools used are available through Volcano Engine’s DataFinder service.