Developed a three-dimensional scoring system to compare 12 car insurance companies in Chile, integrating public data from the CMF and SERNAC.

The model evaluates three dimensions: Territorial coverage (repair shops/10k customers), Repair time (weighted by severity), and Customer Experience (IEC), aggregating them using geometric mean to penalize imbalances.

The analysis revealed counterintuitive findings: the correlation between the number of repair shops and repair speed is practically zero (r=0.03). Small insurers (≤90k customers) outperform large ones in terms of experience (IEC 71.2 vs. 64.8), suggesting that scaling without optimizing operations degrades service.

The result: an actionable benchmark that allows customers to choose according to their priority (speed, coverage, or balance) and insurers to identify competitive gaps.

l.

RESUMEN EJECUTIVO

Problem bussines

The consumer problem

A Chilean who needs to choose car insurance faces more than 22 options without comparable information. Existing rankings use crude metrics (total claims, number of repair shops) that unfairly favor or disadvantage companies based on their size.

The analytical problem

Comparing insurers requires solving three methodological challenges:

Normalization by exposure: a rate of 100 claims in a company with

700,000 customers is better than 50 complaints in a company with 50,000. Metrics should be expressed per 10,000 customers.

Multiple dimensions: consumers value service experience, repair speed, and access to repair shops. These dimensions should be combined without one artificially compensating for another.

Regional variability: workshop coverage varies by region. A national recommendation may be useless for someone in Aysén.

PROBLEMA DE NEGOCIO

METODOLOGÍA - EL EMBUDO

Transformation of absolute metrics to scores 0-100 using market benchmarks (min, average, max).

Normalización

Recolección de Datos

Extraction of public data from CMF and SERNAC: claims, times, repair shops, and customers from 12 insurance companies (2023-2024).

Creation of composite indices: IEC (experience), Time (speed), Coverage (access).

Construcción de Índices

Score Tridimensional

Validación y Dashboard

Aggregation using weighted geometric mean to penalize imbalances.

Consistency verification and interactive dashboard construction in Power BI.

Hallazgos

1. Cobertura ≠ Velocidad

The correlation between number of repair shops and repair time is 0.03 (practically zero). BCI has 437 repair shops and takes 42.5 days for serious accidents. Sudamericana G has 810 repair shops and takes 23.5 days. Operational efficiency matters more than gross coverage..

2. Paradoja de escala

Small insurers (≤90k customers) outperform large ones in customer experience: average IEC 71.2 vs. 64.8, Time Score 83.8 vs. 75.7. Growing your customer base without scaling your operations degrades service.

3. El Gran Diferenciador

Cuando analizamos el Score tiempo

When we analyze the Time Score, the time taken to handle serious claims is the big differentiator. The comparative curve shows peaks in some companies. This

shows that "Serious Days" not only has much higher levels, but also greater variability

Average days

• Mild: 17.84

• Moderate: 25.34

• Severe: 39.26

Dispersion (standard deviation)

• Mild: 2.38

• Moderate: 4.05

• Severe: 5.39

Range (max – min)

• Mild: 6.14

• Moderate: 11.83

• Severe: 15.21

INTERACTIVE DASHBOARD

Dashboard: Three levels of analysis

viPage 1 - Market Overview (Market Benchmark): Visualizations of key findings (distribution of SERNAC results, benchmark of times by severity, coverage-speed ratio). No explanatory text because the graphs speak for themselves.

Page 2 - Competitive Position: Select a company and see how it performs versus the market in each dimension. Identify where it leads (competitive advantage), where it is average (parity), and where it loses (gap to close).

Page 3 - Operational Diagnosis: Delve deeper into the levers that explain the result. If the Time Score is low, it shows whether the problem lies in minor claims,

moderate or severe. If the IEC is low, break it down by complaint rate, SERNAC resolution, or lack of response. Provide specific points of intervention.

CONCLUSIONS AND RECOMMENDATION

📊 CONCLUSIONES

Serious claims define the winners

The Time Score in serious cases shows the greatest dispersion in the market (range 41- 100), while in minor cases the differences are compressed. The insurers that dominate complex cases—Porvenir, Sudamericana, FID with a score of 100—lead the overall ranking. Optimizing the process for serious cases has a higher ROI than improving minor cases.

The "more infrastructure = better service" model does not work.

The correlation between the number of repair shops and repair time is practically zero (r=0.03). BCI has the largest customer base (691k) and an extensive network, but its repair times for serious accidents (42.5 days) are almost double those of competitors with fewer resources. Sudamericana G, with similar coverage, achieves 23.5 days. The efficiency of the network matters more than its size.

Scaling without operating degrades the experience.

Small insurers (≤90k customers) consistently outperform large ones: average IEC 71.2 vs. 64.8, Time Score 83.8 vs. 75.7. Portfolio growth without proportional investment in operational capacity erodes quality. Size is a liability if not managed.

Escalar sin operar degrada la experiencia

Las aseguradoras pequeñas (≤90k clientes) superan sistemáticamente a las grandes: IEC promedio 71.2 vs 64.8, Tiempo Score 83.8 vs 75.7. El crecimiento de cartera sin inversión proporcional en capacidad operativa erosiona la calidad. El tamaño es un pasivo si no se gestiona.

.

🎯 RECOMENDACIONES POR COMPAÑÍA

Suramericana G — Convert coverage into speed Score: 78.4

| IEC: 62.4 | Time: 100 | Customers: 182k

Strength in response times and territorial coverage (Top 3), but customer experience below potential.

• Close RC gap: Eliminate 1.65% of "no response" cases with defined response SLA (≤48 hours) and standardized messages. Target: RC=100.

• Review claims policy: Second review of "not accepted" (45%) with documented criteria and checklist to reduce avoidable rejections. Current rate of 19.6/10k is above average.

• Proactive case management: Milestone notifications, single escalation channel, proactive contact in case of delays.

• Communicate strengths: While the claims front is being corrected, position "fast and with national coverage" — timing is your best asset.

BCI — Urgent operational shock

Score: 56.7 | IEC: 61.1 | Time: 65.6 | Customers: 6G1k

Leader in portfolio size but lagging in experience. Scale became a burden.

• Intervene in serious cases: Assessment ≤48h, authorization ≤72h, proactive follow-up. The serious score (53.4) is 47 points below the leaders.

• Audit workshop network: Measure effective capacity (claims/month per workshop), not just count. Concentrate flow in workshops that meet SLA and weed out those that do not.

those that do not.

• Maintain RC=100: This is the only metric at an optimal level—don't lose it.

• Reduce "no acceptance": 42% rejection rate is high; review criteria and train assessment teams.

Porvenir — Leader to defend

Score: 8G.G | IEC: 100 | Time: G5.0 | Customers: 15k

Privileged position but vulnerable due to scale. Strategy: grow without losing the advantage.

• Leverage differentiation: Communicate IEC=100, lowest claims rate in the market (2.6/10k), and consistent times as verifiable claims in comparators and performance marketing.

• Maintain RC=100 as a brand promise — zero tolerance for "no response."

• Scale selectively: Expand coverage in strategic regions without diluting operational capacity. Prioritize workshop quality over quantity.

• Monitor degradation: Define early warnings if IEC or Time fall as the portfolio grows.

LIMITATIONS AND FUTURE WORK

• Increase temporality (more periods) to strengthen trends.

• Only includes formal complaints; does not capture unreported dissatisfaction

• National averages hide regional variation in times and coverage