The data and methods behind Happily Research

Research by Happily is the research practice of Happily.ai, an employee experience platform that companies use to run daily check-ins, pulse surveys, peer recognition, and feedback. Every study on this site is built from that platform's data.

Most workplace research rests on an annual engagement survey or a one-time panel, a single snapshot of a moving picture. Happily's data is collected differently. Employees respond to short prompts on most working days, year after year. That produces a continuous, longitudinal record of how people feel and behave at work, rather than an end-of-year summary. It is what lets a study measure how stable a habit is across a full year, or what changes in the 90 days before someone resigns.

We publish this research in the open so HR leaders, journalists, and researchers can use it. The sections below set out the dataset, the instruments behind it, how privacy is protected, how the analysis is done, and how to cite any study you find here.

How the data is collected

The data is generated in the flow of work, not in a separate survey exercise. Through the Happily app, employees voluntarily respond to a short daily check-in, periodic pulse surveys, and recognition and feedback exchanges with their colleagues. Most of these take seconds to complete.

Because the same people respond again and again, the result is a set of per-person time series rather than a stack of disconnected snapshots. That structure is what makes the longitudinal questions possible: how a pattern holds or drifts over a year, how one team diverges from another over many months, what a trajectory looks like before a turning point.

The instruments we use

Wherever possible the research uses established, externally validated measures rather than metrics invented in-house. That keeps the results legible and comparable to work done elsewhere.

• WHO-5 Well-being Index. A five-item scale developed by the World Health Organization, validated and used in clinical and academic research worldwide. It lets our well-being numbers be read against external benchmarks.
• eNPS (employee Net Promoter Score). The standard single-question measure of whether employees would recommend their workplace.
• Daily check-in. A short daily prompt on how an employee feels, answered on a five-point scale.
• Peer recognition and feedback. Structured records of who recognizes, and gives feedback to, whom across a team or company.

Where the data comes from

Happily works with companies primarily in Thailand and across Southeast Asia, and the dataset reflects that footprint. We state this plainly for two reasons.

First, every study reports its own sample, population, and time window, so you can see exactly what each finding rests on. Second, the findings are workplace mechanisms: how manager behavior cascades to a team, how recognition spreads, how well-being relates to performance. These mechanisms are not specific to one region, and they are the part of the research that travels. Where a study is limited to a single market or a particular kind of company, its limitations section says so.

How we handle privacy

Research is conducted on de-identified, aggregated data. Individual employees are never identified in any study. Analyses run at the level of groups, teams, companies, or anonymized cohorts, and identifiers are hashed before analysis. Test and internal accounts are excluded. No study reports a result that could single out a specific person.

How we analyze it

Every article on this site is held to the same standards, so a reader can check the work rather than take it on trust:

• Each study ships a methodology box stating its sample, time window, key definitions, and statistical methods.
• We report effect sizes, not only statistical significance, so the size of a finding is visible, not just its existence.
• We distinguish correlation from causation and say which one the design supports.
• We control for the obvious confounds, such as tenure, company, and team composition, and say so when we cannot.
• Every study ends with an honest limitations section. When a result is underpowered or ambiguous, we say so rather than dropping it.

How to cite this research

The studies here are open and meant to be referenced. You are welcome to quote, link, and cite them. We ask only that the work is attributed to Happily Research and that the study is linked.

For a general reference to the research practice, cite Happily Research, the research practice of Happily.ai, at research.happily.ai. For questions about the data or a specific study, get in touch.