Powering better people decisions across the Gulf.
From the evidence to the decision — vendor-neutral research, summarised and sourced, for the people who build organisations.
Latest Decision Briefs
You want recognition to reward the right things. Which criteria can you actually measure, and does recognition tied to them work?
A recognition criterion is measurable when someone else could check it and reach the same answer: a count, a rate against a target set in advance, an observation recorded against a checklist, or a rating from customers the manager did not choose. 'Goes above and beyond' is not one. Whether recognition tied to such criteria works is less settled than the confidence around it suggests. A 2025 review of 43 studies found that most relied on what people said about feeling appreciated: of the 27 quantitative studies, 19 were one-off surveys and one was an experiment. The largest study, of 25,285 employees in South Korea, links recognition to engagement, but its measure of recognition includes pay. In a randomised trial on eight construction sites, sites were recognised only when an audited safety score beat a target fixed in advance, and safety climate improved slightly. The most quoted figures come from companies that sell recognition software. None of this evidence is from the Gulf.
Read→People come back from the course. Does anything change in how they work, and what decides whether it does?
Most organisations judge training by who attended and whether they liked it. The research judges it by transfer: whether people use what they learned. Our evidence is thin and nearly all self-reported: three employee surveys and one employer survey, none of them from the Gulf, plus one review of workplace safety studies that judged results by injury records. Whether people used their training depended on the job they went back to: support from colleagues, feedback, and enough capacity to try it, while heavy demands worked against it. Whether a manager's support matters is unsettled, and the two studies that disagree measured different things: one measured what people did, the other only whether they wanted to. Where outcomes have been measured objectively, in workplace safety, training on its own has rarely moved them.
Read→You have argued about the rating scale, the curve and the cadence. Has anyone asked whether the person being rated believes the process was just?
Most performance-management decisions are about mechanism — the scale, the curve, the cadence, whether AI writes the first draft. The research points at something the mechanism cannot fix. What predicts whether a rating is accepted, and what follows it, is whether the person rated believes the process was just; and the route runs through their satisfaction with the appraisal rather than the score itself. Two practical traps sit underneath that. Reviewing more often is not automatically better — the relationship between interval and proactive behaviour is an inverted U. And at scale, evaluation instruments carry measurable demographic bias. The uncomfortable part for the Gulf is that the survey you would use to check any of this is subject to the same fear it is trying to measure.
Read→What we don’t do
We don’t sell software, run certifications, or publish anything we can’t source. When the evidence is thin the brief says so, and when a study sits behind a paywall we cite it rather than reprint it.
One 5-minute evidence brief, every Monday. Free. No ads. Unsubscribe anytime.