Skorku GEO Dashboard

Benchmark Scope

225 queries across 5 intent categories evaluated across 3 leading foundation models (ChatGPT, Claude, Gemini) to measure brand share of voice (SOV), mention rank propensity, and recommendation distribution across the Indonesian credit scoring category.

• Branded (40 queries): Head-to-head comparisons (Skorku vs Skorlife vs MyIdScore), feature sets, pricing, legal status, and data security inquiries
• Organic (60 queries): Category problem-solving searches without brand bias (cara cek skor kredit online gratis, BI checking, SLIK OJK)
• GSC-Derived (25 queries): Real traffic keywords mined directly from Google Search Console — high-CTR queries and alternative-seeking behaviors (selain idebku, alternatif gratis)
• Long-Tail (50 queries): Real-world emergency cases, edge-cases, approval criteria, and error disputes (Kol 5 bisa pinjam?, komplain data salah)
• Semantic (50 queries): Natural conversational queries mirroring real user prompting in LLMs (Saya mau ajukan KPR... lebih baik ke mana?)

Evaluation Protocol

• System Prompt (Neutral): Jawab objektif dalam bahasa Indonesia, maksimal 250 kata. Koreksi asumsi keliru dalam pertanyaan. Jangan mengarang fakta atau sumber.
• Model IDs (Frozen): ChatGPT (cx/gpt-5.6-sol), Claude (cc/claude-sonnet-5), Gemini (ag/gemini-3.8-flash)
• Temperature: 0.2 (deterministic ranking, eliminating sampling randomness)
• Mention Rank Extraction: #1 = brand whose keyword/alias appears first in response text. #2 / #3 = subsequent mentions. "—" = brand not mentioned
• Data Quality: 675 total executions (100% complete batch). Generative evaluation without search augmentation to measure pure pre-trained brand authority.

Key Findings

• Skorku leads ChatGPT & Claude decisively with 25.8% and 20.0% SOV, maintaining top average rank #1.26–#1.29 (first mentioned). This reflects superior pre-trained authority and domain relevance in OpenAI and Anthropic architectures.
• Skorlife parity in Gemini: Both brands register 25.8% SOV with identical #1.43 average rank. Google's model demonstrates the highest competitive parity for credit management queries.
• MyIdScore underperforming: SOV remains low at 5.3%–15.6% across all models with slower rank positions (1.54–1.83), confirming that legacy Pefindo consumer branding has not translated effectively into generative AI citations.
• Claude conservative baseline: Claude exhibits lower overall brand citation rates (5.3%–20.0%), focusing more on regulatory and neutral educational responses rather than product endorsements.

Strategic Recommendations for Management

1. Defend & Expand ChatGPT/Claude Lead (High Priority) — Skorku's +10–15pp SOV advantage is a major moat. Reinforce this with high-authority topical content on SLIK alternative access, OJK compliance, and debt advisory playbooks.
2. Launch Gemini Intercept Campaign (Q4 Initiative) — Counter Skorlife's parity on Gemini by targeting high-intent Google search queries like "alternatif idebku" and "bi checking online tanpa antre" to train web-indexed RAG embeddings.
3. Deprioritize MyIdScore as Primary Threat — Data shows MyIdScore lacks traction in consumer AI recommendations; allocate digital positioning budgets primarily against Skorlife's consumer expansion.