tiktok-app-performance-tracker

TikTok App Performance Tracker

Purpose

Track, analyze, and optimize TikTok app install campaigns. This skill helps advertisers understand app event metrics, reconcile TikTok data with MMP (Mobile Measurement Partner) data, and navigate iOS 14+ SKAN attribution challenges.

The Critical Landscape: TikTok App Advertising

┌─────────────────────────────────────────────────────────────────────────────┐
│                    TIKTOK APP ATTRIBUTION ECOSYSTEM                          │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│  ANDROID (Clear attribution)          iOS (SKAN-based)                      │
│  ─────────────────────────            ────────────────                      │
│                                                                              │
│  TikTok tracking:                     TikTok tracking:                      │
│  ├─ Device-level attribution         ├─ SKAdNetwork only (post iOS 14.5)   │
│  ├─ Real-time data                   ├─ 24-48 hour delay                   │
│  └─ Full event stream                └─ Limited conversion values          │
│                                                                              │
│  MMP integration:                     MMP integration:                      │
│  ├─ Click + view attribution         ├─ Probabilistic modeling             │
│  ├─ Full user journey                ├─ Limited post-install events        │
│  └─ 100% match expected              └─ 40-60% typical match rate          │
│                                                                              │
│  Discrepancy: 5-15%                   Discrepancy: 20-50%                   │
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

When to Use This Skill

Invoke when user mentions:

• Install tracking: "How many app installs am I getting?"
• CPI analysis: "What is my cost per install?"
• MMP discrepancies: "Why don't TikTok and AppsFlyer match?"
• iOS tracking: "How many iOS installs via SKAN?"
• In-app events: "How many registrations after install?"
• Retention: "What is my D2 retention?"
• Gaming metrics: "How many users complete the tutorial?"
• Subscription apps: "What is my cost per trial?"

Key Metrics Explained

App Install Funnel

┌─────────────────────────────────────────────────────────────────────────────┐
│                      APP INSTALL FUNNEL & METRICS                            │
└─────────────────────────────────────────────────────────────────────────────┘

  IMPRESSIONS → CLICKS → INSTALLS → REGISTRATION → IN-APP EVENTS → PURCHASE
       │           │          │            │               │            │
       ▼           ▼          ▼            ▼               ▼            ▼
     [CPM]      [CPC]      [CPI]       [CPR]         [CPA per        [ROAS]
                           [CVR]    [Install-to-     event]
                                    Reg Rate]

METRICS HIERARCHY:
─────────────────

Level 1: Acquisition
├─ app_install: Total attributed installs
├─ real_time_app_install: Today's installs (updates hourly)
├─ cost_per_install: Spend / installs
└─ install_rate: Installs / clicks × 100

Level 2: Activation
├─ registration: Sign-ups post-install
├─ install_to_registration_rate: Registrations / installs × 100
├─ complete_tutorial: Tutorial completions (gaming)
└─ start_trial: Trial starts (subscription apps)

Level 3: Engagement
├─ view_content: Product/content views
├─ add_to_cart: Cart additions
├─ add_to_wishlist: Wishlist additions
├─ achieve_level: Gaming progression
└─ day_2_retention: D2 return rate

Level 4: Monetization
├─ purchase: In-app purchases
├─ total_purchase_value: Revenue from purchases
├─ subscribe: Paid subscriptions
├─ spend_credit: In-app spending (gaming)
└─ checkout: Checkout completions

SKAN (SKAdNetwork) Metrics - iOS Only

┌─────────────────────────────────────────────────────────────────────────────┐
│                         SKAN ATTRIBUTION EXPLAINED                           │
└─────────────────────────────────────────────────────────────────────────────┘

WHAT IS SKAN?
─────────────
Apple's privacy-preserving attribution framework (standard since iOS 14.5, 2021).
NO device-level tracking. NO real-time data. Limited conversion values.

SKAN METRICS AVAILABLE:
───────────────────────
├─ skan_app_install: iOS installs via SKAdNetwork
├─ skan_conversion: Attributed conversions
├─ skan_purchase: Purchases (delayed)
├─ skan_registration: Registrations (delayed)
├─ skan_result: Combined results
└─ skan_cost_per_result: iOS CPA

KEY LIMITATIONS:
────────────────
┌──────────────────────────────────────────────────────────────────┐
│ Limitation              │ Impact                                 │
├─────────────────────────┼────────────────────────────────────────┤
│ 24-48 hour delay        │ No real-time optimization              │
│ Conversion values (0-63)│ Only 64 possible conversion buckets    │
│ One postback per user   │ Limited post-install event tracking    │
│ No view-through attr.   │ VTA not supported in SKAN              │
│ Campaign cap (100)      │ Max 100 campaigns per account          │
└──────────────────────────────────────────────────────────────────┘

SKAN DATA TIMELINE:
───────────────────
Day 0: User installs app (SKAN timer starts)
Day 0-24h: Conversion value can be updated
Day 1-2: Timer expires, postback sent to ad network
Day 2-3: Data appears in TikTok (24-48h delay after postback)

Total latency: 2-4 days from install to data visibility

CPI Benchmarks by Vertical

Vertical	Android CPI	iOS CPI (SKAN)	Install-to-Reg Rate
Gaming - Casual	€0.50 - €1.50	€1.00 - €3.00	25-40%
Gaming - Midcore	€1.50 - €4.00	€3.00 - €8.00	15-25%
Gaming - Hardcore	€3.00 - €10.00	€6.00 - €15.00	10-20%
E-commerce	€0.80 - €2.50	€1.50 - €4.00	30-50%
Fintech	€2.00 - €6.00	€4.00 - €12.00	20-35%
Food Delivery	€1.00 - €3.00	€2.00 - €5.00	40-60%
Dating	€1.50 - €4.00	€3.00 - €7.00	35-50%
Streaming	€0.80 - €2.00	€1.50 - €4.00	40-55%
Utilities	€0.30 - €1.00	€0.60 - €2.00	45-65%

Regional Multipliers:

• Western Europe: 1.0x (baseline)
• North America: 1.2x - 1.5x
• UK: 1.1x - 1.3x
• Eastern Europe: 0.6x - 0.8x
• Southeast Asia: 0.3x - 0.5x
• LATAM: 0.4x - 0.6x

MMP Integration & Discrepancy Troubleshooting

Supported MMPs

MMP	Integration	Typical Match Rate
AppsFlyer	Full (OneLink)	85-95% Android, 50-70% iOS
Adjust	Full	85-95% Android, 50-70% iOS
Singular	Full	85-95% Android, 50-70% iOS
Branch	Full	80-90% Android, 45-65% iOS
Kochava	Full	80-90% Android, 45-65% iOS

Why TikTok and MMP Data Don't Match

┌─────────────────────────────────────────────────────────────────────────────┐
│                    DISCREPANCY DIAGNOSIS FLOWCHART                           │
└─────────────────────────────────────────────────────────────────────────────┘
                                    │
                    "TikTok shows more installs than MMP"
                                    │
                                    ▼
                         Check time zone settings
                                    │
            ┌───────────────────────┼───────────────────────────┐
            │                       │                           │
            ▼                       ▼                           ▼
      DIFFERENT TZ             SAME TZ                    NOT THE ISSUE
    (Common cause!)        Check attribution window              │
            │                       │                           ▼
            ▼                       ▼                  Check iOS vs Android
      Align to UTC         Compare windows                      │
      in both systems              │               ┌────────────┼────────────┐
                                   │               │                         │
                    ┌──────────────┼──────────────┐│                         │
                    │              │              ││                         │
                    ▼              ▼              ▼▼                         ▼
             TT: 7d click     TT: 1d VTA    MMP has         iOS HEAVY        ANDROID
             MMP: 7d click    MMP: 1d VTA   different       (SKAN issue)     (Should match)
                    │              │        windows                │              │
                    ▼              ▼            │                  ▼              ▼
             Should match    Should match   Align!          Expected gap     Debug:
             (debug further) (debug further)                (20-50% normal)  - Click tracking
                                                                             - Deep links
                                                                             - Postback delays

Common Discrepancy Causes & Solutions

Cause	Impact	Solution
Time zone mismatch	5-20% variance	Align both to UTC or same TZ
Attribution window	10-30% variance	Match windows in both systems
VTA attribution	TikTok higher	MMP may not track all VTA
SKAN delays	iOS 24-48h lag	Compare same date range after lag
Click URL issues	Missing installs in MMP	Test deep link setup
Postback failures	MMP missing events	Check MMP postback logs
Organic vs paid	Attribution overlap	Review MMP deduplication rules
Re-engagement	Counted differently	Align re-attribution windows

Acceptable Discrepancy Ranges

Platform	Acceptable	Investigate	Critical
Android	<10%	10-20%	>20%
iOS (SKAN)	<30%	30-50%	>50%

iOS SKAN Best Practices

SKAN Optimization Strategy

┌─────────────────────────────────────────────────────────────────────────────┐
│                      iOS CAMPAIGN OPTIMIZATION                               │
└─────────────────────────────────────────────────────────────────────────────┘

CONVERSION VALUE SCHEMA (64 buckets: 0-63)
──────────────────────────────────────────

For E-commerce/Revenue Apps:
┌────────────────────────────────────────────────┐
│ Value │ Meaning                                │
├───────┼────────────────────────────────────────┤
│ 0     │ Install only (no post-install event)  │
│ 1-10  │ Registration / Sign-up                │
│ 11-20 │ Add to Cart / Content View            │
│ 21-40 │ Purchase (€0-€50)                     │
│ 41-55 │ Purchase (€50-€200)                   │
│ 56-63 │ High-value purchase (€200+)           │
└────────────────────────────────────────────────┘

For Gaming Apps:
┌────────────────────────────────────────────────┐
│ Value │ Meaning                                │
├───────┼────────────────────────────────────────┤
│ 0     │ Install only                          │
│ 1-5   │ Tutorial start                        │
│ 6-15  │ Tutorial complete                     │
│ 16-30 │ Level 5 achieved                      │
│ 31-45 │ First purchase (IAP)                  │
│ 46-63 │ High-value IAP / Subscription         │
└────────────────────────────────────────────────┘

For Subscription Apps:
┌────────────────────────────────────────────────┐
│ Value │ Meaning                                │
├───────┼────────────────────────────────────────┤
│ 0     │ Install only                          │
│ 1-10  │ Registration                          │
│ 11-25 │ Trial started                         │
│ 26-40 │ Paywall viewed                        │
│ 41-55 │ Trial converted (monthly)             │
│ 56-63 │ Trial converted (annual)              │
└────────────────────────────────────────────────┘

iOS Campaign Setup Recommendations

1. Consolidate campaigns - Max 100 SKAN campaigns per account
2. Higher budgets - Minimum €50/day per ad group for learning
3. Broader targeting - SKAN works better with larger audiences
4. Patience with data - Wait 3-4 days before optimization decisions
5. Use SKAN + MMP together - SKAN for attribution, MMP for incrementality

In-App Event Tracking Setup

Standard Events by App Type

E-COMMERCE APP                      GAMING APP
──────────────                      ──────────
□ app_install                       □ app_install
□ registration                      □ registration
□ view_content                      □ complete_tutorial
□ add_to_cart                       □ achieve_level
□ add_to_wishlist                   □ spend_credit (IAP)
□ checkout                          □ subscribe
□ purchase                          □ purchase (IAP)
□ add_payment_info                  □ day_2_retention

SUBSCRIPTION APP                    FINTECH APP
────────────────                    ───────────
□ app_install                       □ app_install
□ registration                      □ registration
□ start_trial                       □ view_content
□ subscribe                         □ add_payment_info
□ purchase                          □ purchase (deposit)
□ day_2_retention                   □ checkout (KYC complete)
                                    □ subscribe (premium)

Event Value Tracking Best Practices

Event	Track Value?	Why
purchase	YES	ROAS calculation
subscribe	YES	LTV prediction
add_to_cart	Optional	Funnel analysis
spend_credit	YES	Gaming revenue
registration	No	Boolean event
view_content	No	Volume metric

Campaign Optimization Framework

Learning Phase for App Campaigns

┌─────────────────────────────────────────────────────────────────────────────┐
│                      APP CAMPAIGN LEARNING PHASE                             │
└─────────────────────────────────────────────────────────────────────────────┘

REQUIREMENT: 50 conversions in 7 days per ad group

For app installs (CPI €1.50):
├─ Minimum budget: 50 × €1.50 / 7 = €10.71/day
├─ Recommended: €50/day (safety margin)
└─ Conservative: €30-40/day

For in-app events (CPA €15):
├─ Minimum budget: 50 × €15 / 7 = €107/day
├─ Recommended: €150/day
└─ Consider: Start with install optimization, switch after data

LEARNING PHASE STATUS:
─────────────────────
✅ Active: Getting conversions, still learning
⚠️ Limited: Not enough conversions (<50/week)
❌ Not Learning: Stuck - change targeting or bid
✅ Completed: Ready to scale

Campaign Type Selection

Campaign Type	Best For	Notes
Standard App Install	Testing, control	Manual targeting and bidding
Smart+ App	Scaling, automation	TikTok's automated campaign (like Meta Advantage+); algorithm handles audience + creative rotation; needs 5+ ad variants
Value-Based Optimization	ROAS-focused	Requires in-app purchase events

Smart+ App campaigns (part of TikTok's Smart+ suite, expanded Jan 2026) are recommended once you have 50+ installs/week of history. They reduce manual audience management and typically improve CPI by 10-20% after the learning phase.

Bid Strategy Selection

Objective	Strategy	When to Use
Maximize installs	Lowest Cost (auto)	Starting out, broad testing
Target CPI	Cost Cap	Know your target CPI, want control
Value optimization	Value-Based	Have purchase data, optimize ROAS
iOS campaigns	Lowest Cost	SKAN works best with volume
Smart+ App	Auto (algorithm-managed)	Let TikTok optimize bids automatically

Scaling Decision Tree

                         Is CPI below target?
                                │
              ┌─────────────────┼─────────────────┐
              │                 │                 │
              ▼                 ▼                 ▼
            YES             MARGINAL              NO
         (>20% below)       (±10%)            (>10% above)
              │                 │                 │
              ▼                 ▼                 ▼
        Scale up           Hold steady      Troubleshoot:
        (+20-30% budget)   Monitor 3-5 days ├─ Creative fatigue?
              │                 │            ├─ Audience saturation?
              ▼                 ▼            ├─ Competition?
        Check install-     Check retention  └─ Seasonality?
        to-registration         │
        rate                    │
              │                 │
              ▼                 ▼
        Good (>30%)?       Good (>20% D2)?
              │                 │
        YES → Continue     YES → Continue optimizing
        NO → Check LP/app  NO → Check app quality/UX

Output Template

When analyzing TikTok app performance, provide:

## TikTok App Performance Analysis

### Campaign Overview
| Metric | Value | Benchmark | Status |
|--------|-------|-----------|--------|
| Total Installs | X | - | - |
| CPI | €X.XX | €X.XX (vertical avg) | [✅/⚠️/❌] |
| Install Rate | X% | >2% | [status] |
| Real-time Installs (today) | X | - | - |

### Funnel Analysis
| Stage | Count | Rate | Benchmark |
|-------|-------|------|-----------|
| Installs | X | - | - |
| Registrations | X | X% of installs | >30% |
| Add to Cart | X | X% of registrations | >20% |
| Purchase | X | X% of installs | >5% |

### iOS (SKAN) Status
| Metric | Value | Note |
|--------|-------|------|
| SKAN Installs | X | Delayed 24-48h |
| SKAN CPA | €X.XX | - |
| Data Freshness | X days ago | - |

### MMP Reconciliation (if applicable)
| Source | Installs | Variance | Status |
|--------|----------|----------|--------|
| TikTok | X | - | - |
| MMP (AppsFlyer/Adjust) | X | X% | [Acceptable/Investigate] |

**Variance Analysis:**
- [Explanation of any discrepancy]
- [Root cause if identified]

### Recommendations

**Immediate Actions:**
1. [Action based on data]
2. [Action based on benchmarks]

**Optimization Opportunities:**
- [Creative refresh if needed]
- [Targeting adjustments]
- [Bid strategy changes]

### Learning Phase Status
- Status: [Active/Limited/Completed]
- Conversions this week: X/50 required
- Budget recommendation: €X/day

Monitoring Checklist

Daily Checks

• Real-time installs trending vs yesterday
• CPI within acceptable range
• Install rate stable (>2%)
• SKAN data arriving (iOS)
• MMP postback status (if applicable)

Weekly Checks

• Install-to-registration rate
• D2 retention rate
• Purchase/revenue metrics
• TikTok vs MMP reconciliation
• Learning phase status per ad group
• SKAN conversion value distribution

Monthly Review

• CPI trends by vertical/geo
• LTV:CAC ratio analysis
• iOS vs Android performance split
• Creative performance by install rate
• Seasonal patterns identification

MCP Tool Examples

Pull app install and funnel metrics:

# Get install counts and funnel events
tiktok_get_report(
  start_date="2026-03-28",
  end_date="2026-04-04",
  level="campaign",
  metrics=["app_install", "real_time_app_install", "cost_per_install",
           "registration", "purchase", "spend",
           "skan_app_install", "skan_cost_per_result"]
)

# List campaigns to identify app install campaigns
tiktok_query(entity_type="campaigns")

---

Based on 2025-2026 TikTok Ads and MMP integration research. iOS attribution via SKAN requires patience - data delays are normal, not errors.