Skip to main content
A report is a view of a selected set of traffic. Dates define the time window, filters choose which records belong in the analysis, and grouping decides how to divide the result. Sorting changes the order, not which campaign traffic was originally recorded. Set the scope before interpreting a difference. A campaign report for yesterday in UTC and one for yesterday in a different timezone can contain different visits even though both controls say “Yesterday.”

​
Understand which date receives a conversion

Aggregate reports attribute events to the original visit timestamp. Raw Events uses the event’s own timestamp. Consider this flow in UTC:
After processing, the conversion and its revenue belong to September 10 in the aggregate campaign report. The individual purchase appears on September 11 in the raw Events view. This keeps the conversion result with the traffic that produced it. It also means a past day’s revenue can increase when visitors convert later. If you are looking for a specific purchase, start with the raw event’s date. If you are evaluating the traffic that generated it, use the visit date.

​
Set a fair comparison window

  1. Choose the reporting timezone that matches your comparison.
  2. Select the date range and apply it.
  3. Open one campaign or add a campaign filter.
  4. Check the active filter chips and entity status before reading totals.
  5. When comparing with an external system, match both its timezone and its date attribution method.
The report timezone determines the boundaries of calendar days and time groupings. A visitor’s recorded timezone is a separate traffic characteristic. Grouping visitors by that characteristic does not change the report’s date boundary. For hourly patterns, group by Hour to compare hour-of-day buckets. For individual calendar days, use Date. Day of week combines the same weekday across the selected period. Weekday numbering follows Monday as 1 through Sunday as 7.

​
Filter before drilling into a segment

Suppose a campaign has 1,000 visits, but you want to compare mobile traffic from Canada and the United Kingdom. Use a country inclusion filter for those two countries and a device-type filter for mobile. Then group by Offer. The resulting offer rows describe the selected country and device segment. They should not be compared directly with an unfiltered campaign total as if both represented all 1,000 visits. A metric filter changes which result rows qualify. For example, Visits > 100 can remove low-volume segments from the result. Remove it when checking whether all campaign traffic is accounted for. Apply changes after editing the controls. To diagnose an unexpectedly empty result, first remove optional filters rather than changing campaign configuration.

​
Check active and archived entities

Entity reports can filter by active or archived status. If an older campaign disappeared from your comparison, check this status filter. Archiving an entity is different from erasing its historical traffic. Some overview dimensions can include entities or time buckets with zero activity. A zero row tells you there is nothing measured in the current scope. It does not by itself prove that the entity is configured incorrectly.

​
Sort for the question you are answering

Click a supported column header to choose the sorting field and direction. Most reports initially sort by Visits descending. Date, Month, and Day of week use ascending order by default, while Hour starts descending. For the country example, sorting by Profit identifies the largest contribution or loss. Sorting by ROI highlights efficiency, but check the underlying cost and volume before comparing a small segment with a large one. Built-in sortable fields are handled by the report query. Custom metrics sort the rows loaded in the browser. A locally sorted page is not a guaranteed ranking of every row in a larger report. Raw event views use fixed chronological paging with newest records first. Use the raw view’s filters and dates to find a record rather than expecting aggregate sorting behavior there.

​
Explain an apparent mismatch

For a full reconciliation workflow, use Numbers do not match.