Comparison widget

A Comparison widget puts several slices of the same data side by side on a dashboard and marks which slice comes out ahead. You define a few groups, each the same metrics over different filters: one audience against another, a market or segment against the rest, or the two arms of an A/B test. The widget lays them out in a table, a column per group, with the leading value in each row highlighted.

The other widgets tell you how a number is doing; this one tells you which slice is doing it better. Reach for it to benchmark campaigns against each other, confirm which side of a test actually won, or see where a metric is slipping.

It compares groups to each other, not one period to an earlier one. For change over time, that’s the dashboard’s comparison period.

Comparing groups

A group is one slice of the widget’s dataset: a name and a set of filters. Every group draws from the same dataset and the same metrics, and differs only in its filters, so a group is that dataset seen through one condition. Filter Group A to one audience and Group B to another, one market or channel against the next, or a single campaign type against everything else, and the widget holds them next to each other.

Add as many groups as the question needs. They take the names Group A, Group B, Group C, and so on until you rename them. Any global filter on the dashboard applies on top of each group’s own filters, so every group still reflects the same board-wide slice.

How the table reads

The widget is a table: the metrics you chose run down the side, one row each, and every group takes a column. Each row is scored on its own, so one group can lead on opens while another leads on revenue.

A cell reads like this:

• Value: the group’s figure for that metric, over the dashboard’s date range.
• Winner: the best value in the row shows in green, labelled highest (or lowest, where a smaller number is better, as with unsubscribes or spam complaints).
• Gap: every other cell shows in red, with the percentage by which it trails the row’s winner.
• Zero: a group whose value is zero for a metric reads muted, without a percentage.

The colours let you see at a glance where each slice is strong and where it falls behind.

The metric chart

The table gives you the totals; the chart gives you the shape. Click a metric’s row and the table is replaced by a chart of just that metric, each group drawn as its own line across the dashboard’s date range, so you see not only which group won but how the groups pulled apart over the period. Click Back to return to the full table. The chart draws as a line by default; its type, granularity, and scale live in the widget’s chart settings, alongside the rest of its setup.

Set up the widget

Add the widget from the Details section of the widget gallery, where it sits beside the table and list. In its editor:

1. Pick the dataset to draw from.
2. Choose the metrics to compare; each becomes a row.
3. Build the groups: Add group, give it a name, and set the filters that define its slice.

The widget reads the dashboard’s date range, so you don’t set one on the widget itself. Opening and saving the editor work as they do for every widget; see the Dashboards page.