How to Design Dashboards for Multi-Location Organisations

Define how region, organisation, date and assessment status affect the result. Show an empty result as “no completed assessments” when that is what it means; displaying zero would suggest measured performance. Preserve historical questionnaire and scoring versions so a rule change does not silently rewrite an old comparison.

Benchmarks need a stated population: the locations and assessments included in the comparison. Comparing one location with every location in a group may be inappropriate when their operations or assessment criteria differ. We would settle those inclusion rules before offering a benchmark selector, then show enough context for someone reading an exported report to understand the comparison too.

Create an illustrative dataset with two regions, several locations, an incomplete assessment and a not-applicable answer. Calculate expected results independently, then test filtering and export behaviour. Include a user who must not see one of the regions.

For a hypothetical scoring check, give one completed assessment four equally weighted questions: three earn 1 out of 1 and one is not applicable. If the agreed rule excludes not-applicable questions, the expected result is 3 out of 3, or 100%. Counting that answer as zero produces 3 out of 4, or 75%. Neither the chart nor its colour can resolve that difference: the scoring definition must say which calculation applies. Check separately whether the incomplete assessment should contribute at all.

Use these hand-calculated results to check the summary, drill-down and export. A filtered regional view should contain only the assessments permitted by both the filter and the user’s access. Keep the calculation definitions with the report handover so a later change can be checked against the same examples. A smaller set of views that people can explain is more useful than a large collection of unexplained metrics.