We're using FB analytics to check our DAU of paying users only, using a Segment.
When viewing the DAU graph for the period of 1/2017-10/2017, we get a graph of results. But when extending the period to 1/2016-10/2017, the graph is completely different, the values of the same months (1/2017-10/2017) are more then doubled.
When clearing the segment the graph is stable regardless of time period chosen.
Is this a known issue when using segments?
Is there something we can do to fix it?
Is there a way to know which of the graphs is more accurate?
This is expected behavior and it's not a matter of which graph is "more accurate" since you're looking at different segments of users in each case.
In the first case, all users who have made a purchase in 1/2017-10/2017 (10 months) are included in the segment. In the second case, all users who have made a purchase in 1/2016-10/2017 (22 months) are included in the segment.
So if you're looking at the number of DAU on, say, 2/1/2017, in the first case, all the users who used your app that day who made a purchase in a 10-month period are included. Whereas in the first case, all the users who used your app that day who made a purchase in a 22-month period are included. This explains why the numbers of users are more than doubled.
If you just want to see the number of unique users who made a purchase each day, you can do so in the Revenue, Dashboards or Events section by selecting the "purchase" event and "unique users" metric. These numbers will be stable regardless of the date range you select, since no event-based segment is applied.