Whale Tier (Payer Segmentation)

Whale Tier (WhaleTier) automatically tiers users by cumulative top-up amount, identifying whales (big spenders) / dolphins (mid spenders) / minnows (small spenders) / free users — a classic model for game and monetization operations. It corresponds to the backend AnalyticsService.WhaleTier.

1. What Questions It Answers

• After tiering users by cumulative top-ups, what is the headcount share of each tier?
• Which tier contributes the most revenue? (does the classic "whales drive most revenue" pattern hold)
• What is the whales' ARPPU? How many are there?

2. Payer Tiers

Each PayTierSegment contains: tier / tier_label Chinese name / user_count / percentage / total_amount cumulative top-ups of the tier / revenue_share revenue contribution share / arppu the tier's ARPPU.

3. Backend API

gRPC: AnalyticsService.WhaleTier

⚠️ WhaleTier is the only one of the 25 models whose request has no timeRange — it measures users' lifetime cumulative top-ups, not sliced by a time window.

Response WhaleTierResponse: segments[] (each PayTierSegment), total_users, total_revenue.

HTTP (admin proxy)

POST /admin/v1/analytics/whale-tier
Content-Type: application/json

{
  "appId": 0
}

4. Typical Scenarios

Revenue structure diagnosis

Look at revenue_share: a typical game has the whale tier contributing 60-80% of revenue (whale-driven). If non_pay accounts for the vast majority of users while revenue_share concentrates in whale, it is the classic "a few big spenders carry the revenue" structure.

Whale operations

Identify whale-tier users and run dedicated operations for them (VIP service, exclusive events) — churn in this tier hurts revenue the most.

Payment conversion potential

Look at the non_pay → minnow conversion headroom: the free user base is large, and lifting small-amount conversion (first-purchase bundles) is a revenue growth point.

5. Notes and Caveats

• Cumulative basis: tiering is based on users' cumulative top-up total (not within a time window), which is why there is no timeRange.
• Depends on amount: the amount field of top-up events must be reported correctly, otherwise cumulative amounts are distorted.
• Tier thresholds: thresholds per tier are preset by the backend (based on empirical binning); for custom binning use the OLAP Cookbook.
• Difference from Revenue: Revenue Analysis looks at GMV/ARPU trends within a time window; WhaleTier looks at the lifetime tier structure of users.

6. Related Documents

• Revenue Analysis
• Historical LTV
• Data Analyst Getting Started Guide
• Backend API Contract