Before building permit SLA gold metrics, the city wants to understand turnaround-day outliers. What should engineers do?

Before building permit SLA gold metrics, engineers should profile distributions of turnaround days to spot outliers and understand whether proposed SLA thresholds match local reality. Histograms, percentiles, and extreme long-tail cases reveal data issues, legitimate complex permits, and clock-rule problems before KPIs are hardcoded into gold. Distribution insight prevents the city from publishing SLA dashboards that mislabel normal work as failure—or hide true breaches.

Profiling distributions of permit turnaround days using monthly means to spot SLA gold outliers fails by method: monthly means collapse the distribution and hide the extreme cases the analysis must surface, so percentiles and histograms are required to see outliers. Profiling permit turnaround day distributions in gold tables to spot outliers before SLA builds fails by layer: outlier profiling must occur in exploratory analysis before gold is built, not inside gold tables that already encode the KPI. Profiling sampled permit turnaround distributions with preset cutoffs to spot SLA gold outliers fails by bias: samples can miss long-tail extremes and preset cutoffs prejudge what counts as an outlier before the full distribution is seen and validated.

Exam caveat: Profile before KPI construction. Outlier review informs cleansing and SLA policy; it is not the same as profiling permit turnaround day distributions in gold tables to spot outliers before SLA builds.

Operational check: Compute turnaround-day percentiles and list top outliers with business owners. Decide quarantine vs. include rules, then build SLA gold. Compare dashboard breach rates before and after outlier policy and document why profiling distributions using monthly means was not used to spot SLA gold outliers.