Methodology
The most comprehensive employee experience study of the S&P 500 to date
Using Welliba’s AI-powered EXcelerate platform, we analysed more than 25 million data points from over 150,000 public sources to build the most detailed picture of employee experience (EX) across the S&P 500 ever assembled.
Unlike traditional sentiment analysis, this approach goes beyond opinions and anecdotes. It reveals structural, predictive connections between employee experience and long-term business performance.
Learn the methodology behind this research and how it was designed to be rigorous enough for meaningful benchmarking, while remaining intuitive and actionable for business leaders.
Scope & Unit of Analysis
This annual study evaluates organisational health across the full S&P 500 by analysing employee sentiment, culture, and experience using only publicly available data. The unit of analysis is the company. For 2025, Welliba’s AI agents assessed over 25 million data points gathered from more than 150,000 websites to build a comprehensive picture of employee sentiment in the index.
Data Universe & Collection
All inputs are drawn from the public web. Welliba’s EXcelerate technology, a proprietary system of orchestrated AI agents, searches for publicly available employee sentiment concerning each target organisation, then reads, categorises, and sorts this information. The corpus reflects publicly available text data only, collected at scale from a broad set of sites identified by the agents during discovery.
Processing & Modelling
Once gathered, the content is measured against Welliba’s EX model, which comprises 24 factors of employee experience grouped into six areas. The agents convert the public text into factor-level sentiment, producing scores and narratives that enable benchmarking across companies. These factor scores are then ranked from 1 to 24 to identify each company’s top three positive contributors (“Boosters”) and bottom three negative contributors (“Blockers”).
Measures & Comparative Outputs
The report presents overall employee experience scores and factor-level patterns across the S&P 500, highlighting common boosters such as colleagues and direct manager, and the more varied blockers that appear across communication, conditions, and career. Company and cohort views are used to compare organisational “fingerprints,” showing how factor patterns differ across performance groups.
Linking EX to Business Performance
To relate employee experience to market outcomes, companies were stack-ranked by overall EX score. To avoid distortion from extreme growth, the “Magnificent 7” (Alphabet, Amazon, Apple, Meta, Nvidia, Netflix, Tesla) and certain AI-related stocks such as Broadcom were excluded. The top 100 EX-scoring firms were compared with the remainder over a five-year window of total shareholder return, revealing a cumulative 5 percent outperformance for the high-EX group.
Reliability & Reproducibility
Consistency is supported by a single, documented pipeline: the same agents and EX factor model are applied across all companies, and outputs include both scores and explanatory narratives for benchmarking. The analysis is re-runnable on a future snapshot using the same workflow, enabling longitudinal comparison in subsequent annual editions.
Privacy & Ethics
The approach relies solely on information already in the public domain. No private employee data are collected; the system reads publicly available text and reports company-level insights derived from that content.
In essence, the S&P analysis combines at-scale public-web evidence with a structured 24-factor EX model and an agentic AI pipeline to generate company-level scores, identify boosters and blockers, and connect employee experience with long-term shareholder returns.