← Cases
Hong KongMedia and entertainment2024 – ongoing

Game Bear

Machine-learning prediction of long-term ad conversion

Game Bear raised its overall ad conversion rate by 20% to 30% by predicting which ads would convert

WorkingVendor case studyVendor ↗

Problem

Game Bear, a Hong Kong mobile game developer, needed to know which of its ads would deliver conversions over the long term. Waiting for full results meant spending on ads that did not perform.

Approach

Game Bear uses Vertex AI to predict an ad's long-term conversion rate from its performance in the first two to three days after release. The team uses the predictions as a reference and distributes only the ads that are expected to convert better. The company plans to extend AI to game features, including AI-powered non-player characters.

Results

20% to 30%increase in overall ad conversion rate from distributing only ads with higher predicted conversion, per Google CloudVendor case study[1]

Takeaway

Game Bear judges ads on a few days of early data instead of waiting for full results. The company did not say how it measured the conversion gain.

Categories

Business function
Marketing and sales
Innovation type
Process innovation
Evidence
Vendor case study

Sources

Source summaries

[1]Google Cloud7 May 2024

Google Cloud's customer story describes how the Hong Kong game developer Game Bear uses Vertex AI to predict long-term ad conversion from the first days of ad performance. It reports a 20% to 30% rise in overall ad conversion rate; the rest of the page covers networking and analytics. The page carries no date; the Internet Archive first saved it on May 7, 2024.

Full details are in the original sources.

Added 4 Nov 2026 · Updated 4 Nov 2026
Open case page ↗