Understanding our benchmark calculation methodology

We’ve built powerful, reliable benchmarks to help you compare your employees’ experience against industry standards. These benchmarks were created using over 8.4 million responses from 7,500+ companies, and we’ve done the heavy lifting to ensure accuracy, relevance, and fairness.

The Benchmarking approach

We followed a structured, research-backed method:

1. Built on 216 research-backed questions

We created a standard set of 216 questions, mapped across 30 key reporting factors commonly used to measure employee engagement and experience.

Here are some of the major reporting factors we used:

• Leadership
• Work-life balance
• Role clarity
• Communication
• Growth and development
• Recognition
• Manager effectiveness and so much more that you can explore directly in the product.

2. Smart question matching using AI-NLP

Not all companies use the exact same wording. So we used AI-powered Natural Language Processing (NLP) to identify contextually similar questions from the datasets provided by our sources. This ensured that:

• Similar questions were matched to our 216 standard ones
• Responses were mapped accurately, even when the language was different

For example:
If one company asked, “Do you feel your work is valued?” and another asked, “Does your contribution matter here?” — our AI will match both to the Recognition reporting factor.

How we calculated the benchmark scores

Once responses were mapped to the standard questions, we used the same calculation methods found in your ThriveSparrow dashboard.

Favourability score

This is based on the percentage of respondents who selected favorable options on the rating scale. Favorable responses are defined as:

• Top 2 choices on any scale:
  
  
  • 4 & 5 on a 5-point scale
  • 6 & 7 on a 7-point scale
  • 9 & 10 on a 10-point scale

We then calculate the favourability score using this formula:

Favourability Score = (Number of Favorable Responses ÷ Total Responses) × 100

Neutral and unfavourable responses are grouped based on our standard favorability thresholds, ensuring consistency across different surveys.

Percentage score

Each response is converted into a percentage based on the rating scale used — whether it’s a 5-point, 7-point, or 10-point scale. Since companies may use different rating scales, we normalize the responses so that scores remain comparable across all formats. 

For example:

• On a 5-point scale, a 4 = 80%, a 5 = 100%
• On a 7-point scale, a 6 = ~85.71%, a 7 = 100%
• On a 10-point scale, a 9 = 90%, a 10 = 100%
  
  

After normalization, we calculate the average of these percentage values to get the final score for that question. 

This approach ensures that benchmarks are scale-agnostic, making it possible to compare results meaningfully, even if companies use different rating systems.

One of the examples: Benchmarking for the INSURANCE Industry

For the INSURANCE sector:

• We used over 152,000 responses
  
• The same AI-matching and normalization process was followed
• Scores were calculated using the same favourability and percentage methods

This gives you reliable, industry-specific comparisons tailored for insurance companies.

We know benchmarks are powerful only when they’re accurate, comparable, and relevant. With AI, a research-based foundation, and a massive dataset, you can be confident that your scores are being compared against the best.

To explore industry-specific benchmarks or dive deeper into your reports, reach out to your account manager.