How RZR Scaled MiniMax’s Talkie Across Markets

DRIVING AIGC GROWTH: FROM USER VOLUME TO USER VALUE

Talkie: Creative AI Community, developed by MiniMax, is an AI-driven platform focused on content creation and interactive communities. Users can generate creative content, express ideas, and engage socially, forming a highly interactive content ecosystem.

With its strong product experience, Talkie rapidly scaled its user base and established a solid presence in core markets. To further drive user lifetime value (LTV) , MiniMax partnered with RZR, leveraging a retargeting-first strategy to achieve three key objectives:

→ Identify and re-engage high-value users
→ Unlock new global growth markets
→ Scale user base and monetization value simultaneously

FROM VALIDATION TO SCALE: A THREE-PHASE GLOBAL GROWTH ENGINE

Talkie and RZR built a sustainable growth engine through three structured phases:validation → scaling → global expansion, with each step driven by data.

Phase 1: Android Retargeting Foundation

Identify High-Value Users

Target: Re-engage users with demonstrated long-term value.

RZR built a tailored strategy:
→ Pulled contextual signals and behavioral data to assess user value
→ Used machine learning to predict LTV→ Built dynamic audience pools that update in real time, replacing static segments

Results:
→ RZR established a stable retargeting model
→ Cost per Acquisition (CPA) achieved significantly below target
→ ROI met and began exceeding expectations

Phase 1 validated the model and laid the foundation to scale.

Phase 2: Scale with Machine Learning

Grow Value Without Sacrificing Efficiency

With the initial model validated, this campaign moved into scaling—one of the biggest challenges in AIGC growth: maintaining ROI while increasing volume.

RZR’s approach:
→ Strengthened machine learning models to automatically identify high-value users
→ Continuous data science iteration to sharpen model performance

Achieved ramp-up within just three weeks

Results:
→ Rapid scale while holding performance above target
→ D1 ROAS significantly exceeded targets
→ D0 and D7 ROI improved simultaneously, achieving a scalable and sustainable growth curve

Phase 2 proved the model could scale. Phase 3 put it to work globally.

Phase 3: Global Expansion

From One Market to Multi-Region Growth
Once performance stabilized, the focus shifted to global expansion.

RZR’s strategy:
→ Replicated the proven U.S. growth model in new Tier-1 markets
→ Selected high-potential regions: UK, AU, DE, CA, NZ
→ Expanded gradually to ensure controlled growth
→ Adapted strategies to local user behavior and market conditions

Results:
→ Successful expansion into 5 key Tier-1 markets
→ Maintained high-quality user acquisition across regions
→ Strengthened both scale and brand presence

This phase elevated the partnership from performance optimization to strategic global growth.

What Drove the Win

ML-Powered User Value Identification

RZR deployed machine learning models to predict LTV, directing budget toward the users most likely to generate lasting revenue.

Dynamic Audience Mechanism

RZR continuously refreshed audience pools to prevent audience fatigue and hold performance over time.

From Single-Market Success to Global ReplicationL-Powered User Value Identification

RZR ran a structured sequence: Data validation → small-scale testing → scaling— that built a foundation for controlled and repeatable growth.

Full-Stack Execution

Growth, analytics, and creative teams operated as a unified pod, compressing the iteration cycle at every phase.