Speaker
Description
Across the statistical community, there is growing recognition that high-quality outputs depend on high-quality processes. The Office for National Statistics – the UK’s largest independent producer of official statistics – is currently undertaking an ambitious project to strengthen quality by systematically mapping, visualising, and continuously improving the production ecosystem. The aim is to achieve ‘Process Excellence’: creating a transparent and consistent approach that identifies risks, inefficiencies, and opportunities for improvement across the statistical lifecycle.
Robust architecture and process management frameworks are central to this initiative, and constitute the foundations for maintaining quality, coherence, and organisational resilience. There is a strong focus on embedding established standards, to ensure our framework aligns with recognised best practice.
At its core, the initiative operationalises quality by capturing end-to-end workflows, aligned to the Generic Statistical Business Process Model (GSPBM), to convert tacit knowledge into explicit process maps, helping define clear roles, decision points, and quality controls. These artefacts form a coherent production ecosystem that promotes visibility across the lifecycle, supporting evidence-based decision making, and enabling targeted improvements where vulnerabilities or inefficiencies emerge.
Process Excellence has been successfully rolled out to the production of several of ONS’ core statistics, and early outcomes have demonstrated improved accountability, earlier detection of quality risks, and efficiency gains without compromising rigour. Beyond the technical benefits, the initiative has also begun instilling a culture where continuous improvement is proactive and shared.
We will present a replicable model for organisations seeking to institutionalise architecture-driven process management as a pathway to strengthening statistical quality, in a world where trust, transparency, and responsible use of AI are paramount.
