Bar & Column

Sorting > Sort by offers Automatic (dates in time order, else largest first), value up or down, name A to Z or Z to A, data order (the visual header's own sort) and A legend segment: type the segment's name, such as Industrial Pumps, in Segment to sort by and the sites are ordered by that product line, not by the stack total. Stacking order puts the largest or the smallest segment at the base.

Sorting > Let readers change the sort adds a small button that cycles largest first, smallest first, A to Z and data order in reading view, so a reader can re-sort without edit rights. Small multiples share one order computed on the totals, so a site sits in the same place in every panel.

Data labels show the value, the % of category, both in one label or the % of total, with Total labels on stacks and Change % to the previous bar. A label goes inside the bar when it fits and outside when not, never into a neighbour's lane; one that fits nowhere is hidden instead of overlapping. Label background puts a plate behind each label with the text contrast kept at 4.5:1.

Bind a measure to Target and every bar gets a tick, a dot or a line through the bars, with Target labels and Mark bars below target hatching the shortfall. The Analytics pane adds up to three reference lines (a value, average, median, percentile, minimum or maximum) and a band, a linear trend with slope and R², and a Pareto line that marks where 80 % is reached. Difference highlight draws a bracket between any two bars with the change, the per cent and, for dates, the CAGR. Error bars, annotations, icons per category, a Total bar, a logarithmic value axis and an axis break for an outsized bar are free too.

Comparison & variance is the Premium layer. Previous year can come from your own measure in Previous year, from a date column in Date (previous year), where the visual sums the current window (Year to date, Latest full year, Latest quarter or Latest month) and the same days a year earlier, or from a date Category, where it reads the bar one year back. Plan / budget and Forecast take a measure each. Comparison shown as draws them as IBCS ghost bars (previous year grey behind, plan outlined, forecast hatched), overlapping bars, bars beside or markers, also per legend segment in clustered layouts.

Variance charts adds Δ bars and Δ% pins on the same category axis, to the right of bars or above columns, with Second variance chart for two comparisons at once. Colour bars by met / missed and Lower is better (costs) colour the result; Category axis > Rank change vs previous year puts ▲3, ▼1 or new next to each category; Sort by > Variance to comparison sorts by the gap; and Narrative > Findings panel (premium) states the total change, the categories that drove it, how many are below plan, concentration, outliers and the biggest movers, each with how it was calculated.

The previous-year window follows calendar-day arithmetic: 29 February maps to 28 February, and a window that ends on the last day of February covers the whole February a year earlier, the same month-end rule as DAX DATEADD. The anchor is the latest date in the data, so every category is cut at the same day. A category with a previous-year value and no actual is kept, because a lost customer is a finding; one with no previous-year value gets no ghost bar and ranks as new.

Blank plus blank stays blank, never zero, and a variance % against a blank or zero base is left empty rather than shown as infinite. Layout > Hide zero and blank categories removes empty categories when you want them gone. Ranks, Other, totals and every statistic are computed the same way for free and Premium users; the numbers are golden-tested against numpy, scipy and pandas.