The Field Guide › Article
Argues that making work visible — mapping WIP (work in progress), dependencies, competing priorities, unplanned work and stale tasks, following Dominica DeGrandis's Making Work Visible — is a precondition for psychological safety rather than a separate concern, because visibility is what lets people say no with any credibility. Too much WIP is identified as the single most corrosive factor, destroying individual and team effectiveness quickly and directly reprising the utilisation/wait-time relationship covered the previous week: a WIP limit only works once the work behind it is actually visible. Illustrates the point with a personal kanban practice — three workstreams, colour-coded task types, physical cards preferred over SaaS boards for the satisfaction of throwing a finished card in the bin — before turning to DeGrandis's four reasons people say yes to work they shouldn't: not wanting to be seen letting the team down, preferring shiny new work to the unglamorous backlog, underestimating how long other people's work takes, and the basic difficulty of saying no to people we like and respect. Psychological safety is what makes candour about capacity possible — being able to say 'no, not now' or 'yes, but X will have to wait' — but that candour still depends on the work being visible in the first place, and visibility systems need to be passive and low-friction, particularly in highly responsive roles such as an emergency department, or they become one more task competing for attention. Draws on firsthand experience leading an engineering department that had drifted into permanent firefighting: heroic, unplanned work that looked busy and felt valuable but delivered little, fixed by visualising what was actually being done (source, duration, cause), discarding low-value work, and identifying where unplanned work originated — which, combined with building the psychological safety needed for the conversations this required, produced a genuinely calm team: better flow, more finishing relative to starting, more slack for learning, and real incidents that were easier to spot and learn from precisely because they weren't buried in background noise. The central tension is that a chaotic team looks busy and a calm one doesn't, illustrated by having to defend to a CEO why a team with a constantly-in-use pool table was outperforming its previous, visibly frantic self — a defence only visualised work data could make. Closes by extending the argument to incentives: a team measured on utilisation, billability, or tasks completed is being incentivised toward chaos and overload, and team metrics need to reward calm, sustainable throughput and visible outcomes instead.