Analytics - helpful to know
We try to make the metrics we present in the analytics dashboard as straightforward as possible, but they still depend on details that may be helpful for you to know. Here are a few things you should be aware of if you try to understand exactly how we come up with a certain number and the behavior of the dashboard.
Analytics refresh (Cache)
There are many things that can cause the analytics results to change, even in the past. For example, if a customer replies before the inactivity time ends to a conversation that was marked as resolved, the conversation will become unresolved.
Analytics are refreshed every hour or whenever a new query is run. For example:
1. If you go to analytics right now, it will pull the latest data
2. If you revisit the same report anytime within the next 1 hour, you will see the same cached data
3. If you revisit the same report anytime after 1 hour, the analytics report will be refreshed
4. If you revisit the same report but add a new criterion (ex. include this new channel), analytics will then recalculate
Analytics are not fixed
Metrics displayed are not fixed, meaning that the analytics can change from one time to another due to the granularity of the segments (duration). So for example, if the inactivity period is set to 5 days and you are watching analytics for the current week, the analytics are going to fluctuate until the selected period is at least past by 5 days.
For example, in the Conversations tab, if you tagged a conversation after a selected time frame of the report, the time frame selected will still reflect the tag in the tag table.
The analytics track each teammate who replies in a conversation, regardless of the official assignee of the conversation. If a conversation is assigned to Mary, but Sophia responds because she is unavailable, Sophia will get a reaction and a response time.
If you modify a conversation, it will usually affect the analytics as expected: deleted conversations will not show in analytics; moving or tagging a conversation will affect views filtering on these inboxes. However, there are 2 exceptions:
- As stated earlier, it may take up to 1h for the results to be updated.
- If the conversation has become inactive, only the latest segment will be updated.
Paul (customer) chats with you to report a problem. You tag the conversation as #support.
- A month later, Paul chats with you again: he is so in love with your product that he would now like to apply for a job. You tag with #job.
In this case, analytics from the previous month will not be modified.
Let's say you reply on Tuesday to a customer email received the previous Friday. If you later check the analytics for this week, we will correctly report the response time, even though the original message is outside of your requested time range.
Auto-replies are ignored by the analytics: they do not count as a reply and do not generate a response time.
When sending a sequence out of Front using a team email address, analytics will count each individual outbound message towards the team inbox's New Conversations metric, which also rolls up into the Total Conversations metric in the conversations view.
Items that are marked as spam in Front or arrive to Front marked as spam are excluded from analytics.
If you are out-of-office, any replies that customers send in conversations that are assigned to you will automatically be unassigned and return to the original shared inbox they came from. Thus, you will not personally accrue response or reaction times because the conversation is not assigned to you anymore.
However, teammates can still manually assign conversations to you while you're out-of-office, which does accrue response and reaction times.
Moves between Teams
When you move conversations from Team A to Team B, analytics are computed for the time that they were in a certain Team. In other words, all analytics are computed for Team A while the conversation is in Team A; and after the move to Team B, all analytics are computed for Team B.