Know yourretention risk beforeyour best people leave.
8 factors3 minutescited research
Not scored
Every finding shows its source and boundary.
Take the diagnosticFactorManager qualityGrowthPayWorkload
Finding
Source
Boundary
Evidence in index
Open the research indexResearch evidence
Can employee turnover really be predicted?What four attrition-model studies show about discrimination, class imbalance, explainability and the gap between a model and a retention policy.Ribes, Touahri and Perthame (2017), preprint4 sources→People analytics needs more than consentThree peer-reviewed studies show why employee benefit, understandable data use, audit trails and recourse belong in people analytics design.Zieglmeier, Gierlich-Joas and Pretschner (2023), peer reviewed3 sources→What turnover research finds before an employee leavesA careful synthesis of tenure, role stagnation, peer departures and job embeddedness, with the context and limitations kept intact.Ribes, Touahri and Perthame (2017), preprint3 sources→Where employee attrition models breakThe recurring limits behind attrition-model claims: synthetic data, single-company samples, imbalance, leakage, explanations without causality and weak employee governance.Ribes, Touahri and Perthame (2017), preprint6 sources→
Open the full research index Where the numbers came from
See all tracesWhere the 33% of salary turnover cost comes fromThe most quoted figure in employee retention is a modelling assumption, not a research finding. We followed it to the document it came from.Professional Advantage2 sources→Three disengagement figures, pointing the wrong wayThe 37%, 49% and 60% disengagement figures are credited to Queen's and Gallup. Gallup's nearest numbers run in the opposite direction, and Queen's published no such figures at all.Gallup, 2012 Q12 Meta-Analysis, Summary of Findings. Gallup's own address for the meta-analysis now returns 404; the copy that resolves is mirrored at employeeengagement.com5 sources→Gallup's 70% manager claim is real. Almost every retelling of it is not.The claim is Gallup's own and their wording is careful. The method behind it has never been published, and most restatements widen it past what it supports.Gallup1 source→
Does it work? What the evidence says
All evidence reviewsDo Stay Interviews Work? What the Evidence SaysStay interviews are widely recommended and rarely tested. The honest case for them rests on the failure of exit interviews, not on direct evidence.Evidence review1 source→Do Exit Interviews Work? What the Evidence SaysWe found no controlled study showing exit interviews reduce turnover. What the research does document is why the instrument underperforms.Evidence review1 source→Do Employee Engagement Surveys Work?The evidence links engagement to lower turnover. It does not show that running a survey lowers turnover. Those are different claims.Evidence review1 source→Do Retention Bonuses Work? What the Evidence SaysReal causal studies exist and they disagree. Effects are conditional, modest when present, and tend to stop when the payments stop.Evidence review1 source→Do Onboarding Programs Work? What the Evidence SaysThe best causal evidence in employee retention is about onboarding. It is one experiment, at one firm, in a call center losing half its staff a year.Evidence review2 sources→Do Employee Recognition Programs Work?The causal evidence on recognition is genuinely good. It measures performance in a three-hour task, and continued activity among unpaid Wikipedia volunteers.Evidence review2 sources→90-Day New Hire Check-In Questions, Built on the Onboarding EvidenceCheck-in questions for a new hire's first ninety days, tied to the three mechanisms the onboarding research finds carry retention: role clarity, self-efficacy and social acceptance. Free to read; download with your email.Question set3 sources→Exit Interview Questions That Get Past the Safe AnswerAn exit interview template built around the two documented ways exit interviews fail, with each question tied to a turnover factor and its source. Free to read; download it with your email.Question set1 source→Stay Interview Questions Grounded in What Predicts TurnoverEighteen stay interview questions, each tied to a factor the research finds predictive, with the citation next to it. Free to read; download the pack with your email.Question set3 sources→
A diagnostic you can inspect.
Answer the factors, see the signal, then read the research behind the result. Nothing disappears behind a score.
- 01Describe the team
Nine focused questions establish the conditions around retention risk.
- 02Read the signal
The result identifies where the pressure appears strongest.
- 03Check the evidence
Sources and boundaries stay attached to the findings they support.
Find the part of the system that needs attention.
Factor. Finding. Source. Boundary. Then a decision you can defend.
Take the diagnostic
