Outliers

Last updated on Oct 07, 2026

Overview

Outliers is a Juno AI agent. It watches daily conversions and return on ad spend for each media source and flags the days a figure moved sharply. Use it to catch a sudden drop or spike within days of it happening, then decide whether the cause is a real change in performance, a tracking problem, or a planned event such as a sale.

Navigation: Juno > AI Agents > Outliers

You also reach this page from the banner at the top of the screen, which reads "We have detected [number] outliers in last 7 days". Select Click here to view.

An outlier is a prompt to investigate, not a conclusion. Confirm the cause before you report an outlier to a client or change spend because of one.

Use it to check

Question Where to look
Did conversions or return on ad spend move sharply on any channel this week? Outliers table
Which media source moved? Media Source column
How large was the move? Changes(%) column
Was the move a one-day event or part of a longer shift? Outliers Trend chart, set to the row's source and metric
How far did the figure land from its usual level? Outliers Trend chart, Value against Mean Value and the two thresholds
Did anyone on the team already review this outlier? Tag column

The toolbar

Control What it tells you
Date range selector Top right of the page. Sets the period to review, such as Last 7 Days
Search Filters the table by the text you enter
Add Tag Tags outliers from the toolbar
Export Exports the table

The Outlier Activity table

Each row is one flagged day for one metric on one media source.

Column What it tells you
Checkbox Selects the row
Date The day the figure moved
Media Source The channel the figure belongs to, such as Google Ads, Facebook Ads, Email, or SMS
Type The metric flagged, such as Conversion or ROAS
Changes(%) The size of the move as a percentage
Value The figure recorded on the day. A count of conversions for Conversion, a return on ad spend ratio for ROAS
Tag Your team's label for the row. Type a tag and press Enter
Action An X icon on each row

Use Items per page and the arrows below the table to move through the results.

The Outliers Trend chart

The chart below the table plots one metric for one media source over time. Pick the pair with the Source and Metrics selectors above the chart.

Element What it tells you
Source The media source the chart plots
Metrics The metric the chart plots, such as Conversion
Value The daily figure, drawn as a shaded area
Mean Value The average level for the metric
Upper Threshold The top of the normal range
Lower Threshold The bottom of the normal range
Up triangle Marks a Deviation Percentage Up
Down triangle Marks a Deviation Percentage Down

Reading an outlier

Match the row to the chart, then read the pattern. Each reading below is a lead for your own judgment, not a diagnosis.

What you see What to consider
Value rises above Upper Threshold for one day, then returns A one-off spike. Look for a promotion, an email or SMS send, or duplicate tracking on the day
Value falls below Lower Threshold for one day, then returns A one-off dip. Look for a site outage, a paused campaign, or a gap in tracking
Value stays outside the thresholds for several days A lasting shift. Look for a budget, bid, audience, or creative change on the source
Conversion and ROAS both fall for one source on the same day A problem on the one channel. Compare its spend and clicks for the day against the platform's own reporting
Conversion falls on several sources on the same day Something site-wide, such as checkout or tracking, more likely than a problem on each channel
The date carries a note for a launch, sale, or outage The move has a known cause. Tag the row and move on

How to use Outliers to make decisions

  1. Open Outliers from the banner or from Juno > AI Agents > Outliers, and set the date range.

  2. Start with the largest Changes(%) values on the sources carrying the most spend.

  3. Set Source and Metrics on the Outliers Trend chart to the row's Media Source and Type. Check whether the move lasted one day or several.

  4. Select View Note in the top bar and check for a note on the same date.

  5. Compare the day in the advertising platform itself. A conversion drop with steady spend and clicks points at tracking. A drop in spend points at budget or delivery.

  6. Tag the row with the cause you found, such as "promo" or "tracking gap", so the rest of the team sees it was reviewed.

  7. Change budgets or campaigns only once the cause is clear. One day outside the normal range is a weak reason to move spend on its own.

What to check and when

When What to check
Each morning The banner count, and any new rows since yesterday
Weekly Every untagged row from the last 7 days
Before a launch, sale, or budget change Add a note for the date, so the outlier it causes reads as explained
After a site, checkout, or tracking change Conversion rows across every source for the days after the change
Before a client report Every outlier in the reporting period carries a tag with its cause

Troubleshooting

Symptom Likely cause Fix
An outlier falls on the day of a planned sale or launch The event moved the figure outside its usual range Expected. No action. Tag the row and add a note for the date
Conversion outliers on several sources on the same day A site-wide checkout or tracking issue Check the site and tracking for the date before changing any campaign
A tag does not stay on the row The tag was typed but not submitted Press Enter after typing the tag
The trend chart shows a different source from the row you are reviewing The Source selector still holds the previous choice Set Source and Metrics to the row's Media Source and Type

Common questions

How do I see the trend behind one row?

Set Source to the row's Media Source and Metrics to its Type in the Outliers Trend chart.

How do I mark an outlier as reviewed?

Type a tag in the row's Tag field, such as the cause you found, and press Enter.

Should I change a budget because of an outlier?

Not on the outlier alone. Check the trend chart, your notes, and the advertising platform first, and act once you know the cause.

Does tagging an outlier remove it from the table?

No. The row stays in the Outlier Activity table with your tag in the Tag column.

Related articles