Compare splits one metric into several series, one per segment, account, campaign, tactic, or dimension. It can also draw the previous period alongside the current one.

Turning on Show as uplift collapses exactly two of those series into a single line that plots the gap between them. Use it when the gap is what you are tracking, like whether Pro plan users adopt a feature more often than Starter plan users.

Compare is available on Metrics cards only.

What you can compare by

Segments, Accounts, Campaigns, and Tactics all work the same way. Add up to five items and the card draws one series each. Uplift narrows that to two.

Dimensions breaks one card down by Page, Domain, Platform, Locale, OS, or Browser, drawing one series per value.

Locale, OS, and Browser add a second dropdown for narrowing to specific values. Leave it on All locales, All operating systems, or All browsers to break the dimension out in full. Page, Domain, and Platform have no value list.

A seventh option, User, appears only on the Sessions metrics and on Count of Events and Per session. On every other metric each series would hold a single user, so it is left out of the list.

Platform, OS, and Browser are recorded only on events from the Appcues SDKs, so a breakdown by any of the three leaves out events sent another way.

Previous period draws the same metric over the window before your current one. 

Build the comparison

1. Open a Metrics card in the card builder and choose a Metric. See Add cards to a dashboard.
2. Under Events, set the events the metric applies to. Conversion rate gives you two slots, Conversion event: over Out of:, with a ÷ between them.
3. Turn on the Compare toggle. The comparison settings appear below it.
4. Choose what to compare by, then add the items you want. Click the X beside an item to remove it; the Add button disappears at the cap.
5. Click Save card.

Compare to the previous period

The Previous period window matches the length of your current date range and ends where that range begins, so a dashboard set to Last 30 days compares against the 30 days before it. Change the date range and the comparison window moves with it.

Single value cards accept only Previous period.

How it draws depends on the card's Analyze as setting:

• Trend over time and Breakdown draw a second series, with the legend Previous period against This period.
• Single value shows the current number with an arrow and the percent change beneath it. Set Show as uplift and the uplift becomes the number instead.

With Show as uplift on, the baseline is always the earlier window, so a positive value means the current period is running ahead of it.

Collapse two items into an uplift line

Uplift works with any metric. Conversion rate is the usual choice, since a gap means more as a rate than as a raw count.

1. Set Show as uplift, at the top of the Compare panel, to % for the relative difference or # for the absolute one. The chart switches to a single line, and the item rows are replaced by two labeled fields.
2. Pick the group you are measuring in Selected.
3. Leave Compared to on Everyone, or pick a second group to measure against. Previous period has nothing to choose.
4. Click Save card.

Either field may be left empty, and an empty field runs unfiltered. Hovering a point shows Uplift (%) or Uplift (#) with the value for that date.

Read the uplift line

Take one card: Conversion rate, with a Report Exported conversion event over All users, Pro plan in Selected and Starter plan in Compared to.

Appcues works out each segment's conversion rate separately for each day, then plots the gap as a single line. Say Pro plan converted at 30% one day and Starter plan at 10%. Compared to holds the baseline, the value you measure from; Selected holds the group measured against it.

What the two uplift units show

# plots the gap between the two rates. % plots that gap as a share of the baseline.

On a Conversion rate card both display as percentages, which is the easy thing to misread. With Pro plan at 30% and Starter plan at 10%, # reads 20%, meaning 20 percentage points, and % reads 200%, because the gap is twice Starter plan's own rate. For a count metric like Unique count of Users, # is the plain difference in the count.

Swap the two fields and the sign flips. A negative value means the selected group converted less often than the baseline, not that anything fell over time. With % the floor is -100%, where the selected group's rate is zero.

What All users counts

An empty Out of: slot divides by the whole basis, shown as All users, All accounts, or All sessions depending on which Conversion rate you picked. That means every user with at least one event in the dashboard's date range, not everyone in your product: anyone with no events in the window is left out. The same rule sets each segment's per-day denominator: the segment's users who had any event that day, not everyone in the segment.

To divide by users who started a session, set Out of: to Session Started. Appcues' own active-user count is broader, because it also counts screen and page views, so the two figures will not match.

What a comparison doesn't tell you

A comparison describes a difference between two groups. It does not tell you what caused the difference, because the two groups differ in more than the thing you are measuring. Pro plan users pay more and use the product more, so they were already likelier to export a report.

Two scoping details change what the line measures:

• Everyone includes your selected group. Leaving Compared to on Everyone runs the baseline unfiltered, so the selected group is counted inside its own baseline. That pulls the uplift toward zero.
• Dashboard filters land on one side only. A card's own filter wins over a dashboard filter on the same dimension. With Pro plan in Selected and a dashboard Segment filter active, the selected side ignores that filter while the Everyone baseline obeys it, leaving the two sides scoped differently.

When the number looks wrong

Check both groups' own conversion rates for that day before reading anything into the number.

Because % divides by the baseline, it spikes whenever the baseline is near zero. A baseline segment that converted one user out of four hundred reads as 0.25%, and a selected group converting at 2.7% against it reads as 980%. A small selected group does the same thing from the other side: a few hundred users means two or three conversions on a quiet day move its own rate enough to push the percentage into the hundreds. Compare a small segment against a large one and expect a spiky line.

If the baseline is exactly zero, Appcues drops that point rather than dividing by it, so the line shows a gap. A card whose every point has a zero baseline draws nothing at all.

Two ways to tell the difference:

• Switch Show as uplift to # for the absolute gap. A large % next to a small # means small numbers on one side or the other.
• Clear Show as uplift to plot both groups separately and look at the baseline directly.

Day-to-day movement on small segments is mostly noise. The card shows no confidence interval, so no single day's value is evidence on its own. Widen the date range, or read the trend across the whole window.

If you don't see Show as uplift

• Confirm the Compare toggle is on, and that you are not comparing by Dimensions.
• On a Trend over time card set to Previous period, two or more events with Match on Each already spend the card's series one per event, leaving none for an uplift line. Set Match to All or Any.
• Picking a segment, account, campaign, or tactic comparison moves Match off Each for you and disables the Each pill while Compare is on. Its tooltip reads "Turn Compare off to chart each event separately."

If a card shows no data at all, see Analytics dashboards overview.