How do you measure the real-world impact of research?

Citations and altmetrics mostly measure academic attention, which is a narrow and lagging slice of impact; the returns that matter — social, clinical, economic, policy — are only partly visible in reference lists (Bornmann 2013). Even large national exercises struggle to capture wider impact with metrics: reviews of the evidence find only a couple of quantitative signals (like patent and clinical-guideline citations) that plausibly reflect societal benefit, and warn against over-relying on any of them (Wilsdon et al. 2015; Hicks et al. 2015). Nabu's Impact potential framework structures this into four readable dimensions: Relevance (is there real demand and uptake?), Readiness (can it be put to use?), Generalisability (does it generalise beyond its original setting?), and Trajectory (does it build on and feed a durable line of work?). Crucially, impact potential is scored separately from quality, because a paper can matter enormously and still be flawed. See the methodology for how the dimensions are calibrated.