Customize table display
Tables make it easy to view the exact values of your analysis metrics, and you can customize them as needed.
Table style
The default table style is Date > Ascending: results for different dates are shown in order from earliest to latest, and each date is a column in the table. To see the most recent dates first, change it to Date > Descending.
In addition to Date, you can also view data by Events. Rows and columns are then transposed: each analysis metric is a column in the table, and each row shows the data for a different date.
If you use group-by items or an event breakdown, you can also select Groups to show the group values as table headers. By default, Groups sorts in descending order by the sum of the first analysis metric. You can switch to ascending or descending lexicographic order of the group values.
When a table in the Events or Groups style is shown on a dashboard, you can also select Break down by date(s) to show dates outside the table, so you can quickly switch dates and compare how different groups perform.
Table display modes and multi-column sorting
For easier reading, tables use Merge Rows headers by default, and you can sort by a specific column.
You can also switch to the Tile display, in which data is sorted by multiple columns in the order you click the headers.
If you selected group-by items when building analysis metrics, you can use the Layered view, in which all group-by items can be expanded or collapsed level by level.
After you expand some group-by items, you can also sort data by multiple columns in the order you click the headers.
The Layered view is unavailable in the following cases: a retention metric is used in the analysis metrics, or an event breakdown is used.
Overview
When the analysis granularity is not Total, Events Analysis automatically calculates an overview for each analysis metric. Click Overview Settings to choose a suitable summary method for each analysis metric. Once configured, in addition to the data for each date, the table also shows the overview result calculated from the data of each date.
If you don't need the overview, clear Show Overview.
The summary methods are:
- Sum: For total count or sum metrics. Simply adds up the results of each period, without deduplication
- Average (rounded): For averaging sum metrics. Calculates the arithmetic mean of the results of each period (and rounds it)
- Weighted Average: For ratio or per-user metrics. Adds up the numerators and denominators of each period separately and then calculates the result. The larger the denominator, the greater the weight of that period
- Max.: The maximum of the data across dates
- Min.: The minimum of the data across dates
- Total: For deduplicated metrics. Deduplicates across the selected time range (including weeks and months that are filled in when you view by week or by month). This is resource-intensive, so use it with care
Compared with the average, the weighted average better reflects the overall picture. For example, here is the data for the 1st~3rd:
| 1st | 2nd | 3rd | |
|---|---|---|---|
| Active Users | 100 | 200 | 500 |
| Paying Users | 10 | 30 | 100 |
| Payment Rate (Paying Users/Active Users) | 10% | 15% | 20% |
Average for the 1st~3rd = (10%+15%+20%)/3 = 15%
Weighted average for the 1st~3rd = (10+30+100)/(100+200+500) = 17.5%
The weighted average can also be understood as:
10% * 100/(100+200+500) + 15% * 200/(100+200+500) + 20% * 500/(100+200+500)
You can also use Exclude Partial Dates to get more accurate overview results, because the data for those dates doesn't reflect actual performance. For example, if you view data at 12:15 on January 1, 2023, with the time granularity set to Daily, the 1st is a partial date. With Hour, 12:00, 13:00, ... 23:00 are all partial dates.
Note that if you use a created retention metric, partial dates follow the same logic as the overview values in the Retention Analysis model. If you view data on the 3rd, the Day 0 retained users on the 3rd, the next-day retention rate on the 2nd and 3rd, and the 3-day LTV on the 1st, 2nd, and 3rd are all treated as partial-date data and excluded from the overview.