Guide section 8 of 31

Table & Matrix Standard

The reusable standard should be a configuration of native Table and Matrix, not a new custom table control.

Table

Use Table for flat detailed records where users need exact values, sorting, scanning, conditional formatting, links/images where supported, tooltips, cross-filtering, export and scrolling.

Matrix

Use Matrix for hierarchical or multidimensional analysis: row/column hierarchies, drill, expand/collapse, stepped layout, subtotals/grand totals and conditional formatting.

TopicStandard
HierarchyTable: flat. Matrix: explicit hierarchy levels; show current depth and expand/collapse affordance.
SortingSet a task-oriented default: rank, chronology or business order. Do not rely on accidental source order.
Totals/subtotalsUse only when meaningful; visually differentiate without overpowering detail.
Row headersKeep labels concise and readable; preserve hierarchy indentation where used.
Expandable hierarchiesPrefer Matrix when expand/collapse is part of the task; avoid simulating it in separate card rows.
Conditional formattingUse semantic color, icons or data bars to aid scanning; always retain a textual/numeric cue.
Data barsUseful for magnitude comparison in dense tables; ensure the numeric value remains readable.
IconsUse as redundant status cues, not as the only carrier of meaning.
SparklinesUseful as compact trend context when supported/governed; do not overload a wide table.
DrillSpecify whether hierarchy drill is allowed and what the default level is.
Column alignmentText left; numeric columns usually right; headers aligned consistently with content.
Column sizingPrioritize key columns; avoid forcing all columns into one viewport if readability collapses.
Numeric formattingStandardize currency, percentages, percentage points, decimal places and abbreviations.
ScrollingTreat scrolling as normal native behavior; avoid nested page/visual scroll where it harms orientation.
Pagination/virtual scrollingDo not design a web-style pager unless the chosen Power BI mechanism actually provides it; test large-result behavior in runtime.
Empty statesExplain no-data vs no-results-after-filter; provide a recovery path such as clearing filters.
Selected statesSelection must be visible and its cross-filter/highlight consequence predictable.
Comparison columnsGroup current/target/previous/variance logically and keep the comparison basis explicit.
VarianceUse sign, unit and semantic status consistently; distinguish % from percentage points.
AccessibilityUse clear headers, logical reading/tab order, non-color cues and consistent formatting.

Recommended reusable variants

  1. Detail table — flat records, sortable, compact density, clear primary identifier.
  2. Comparison table — actual + target/budget + variance + status.
  3. Exception table — prioritized records with severity/reason and drillthrough.
  4. Hierarchy matrix — expandable organizational/product hierarchy with subtotals.
  5. KPI matrix — measures across entities/periods with controlled conditional formatting.
  6. Trend-in-table variant — compact sparkline only when it improves scanning and remains performant.