This guide will cover the most commonly seen kinds of discrepancies seen when querying Facebook Insights. Unless specified in the section, all guidance within should apply to all Supermetrics products that support the Facebook Insights connector.
Quick links to each section:
- Total Reach Doesn't Match Sum of Organic and Paid Reach
- Organic Reach is Missing Values Before October 2017
- Reach (Or Other Calculated Unique Metric) Does Not Match Insights UI/Export Report
- Data Studio Discrepancies with Filters
- Grand totals are calculated from Page totals
Total Reach Doesn't Match Sum of Organic and Paid Reach
This is written into the Graph API as a known issue under the sections for Page and Page Post Impressions (see here for the reference).
Note that the unique impression insights values are calculated independently. Total page/post reach may not always be exactly equal to the sum of paid and non-paid unique values.
As this is how the data comes from the Facebook API, there's nothing we can do to correct it.
Organic Reach is Missing Values Before October 2017
There appears to be an issue with getting some values after October 2017 from Facebook. This affects both the API results and the export report. We aren't clear at this point if this is due to a policy change at Facebook to no longer track these or if it is a bug in their system. Please reach out to Facebook Support with your query and export data so that they can investigate.
Example from Insights Export report for Page data:
Example from Supermetrics query:
Reach (or Other Calculated Unique Metric) Does Not Match Insights UI/Export Report
Unique metrics in Facebook Insights, the most commonly used being "Reach", only work when calculated on 1-day, 7-day, and 28-day date ranges. This is so that Facebook can calculate out the non-unique users within that period. That means though that if you split "Total Reach" by day for 7 days and sum those results, they will not match the week (7-day) result! This is because the values split by day are only being calculated as 1-day values, so if a user came back on multiple days, they will be counted multiple times in the results. They would be removed in the weekly calculation.
Here's an example showing how comparing summed 1-day Reach values to a last 7 days combined Reach ends up looking incorrect and what you should actually be checking against in the Insights report to validate the value.
Facebook Insights data export:
So make sure that:
- You are not summing daily values to compare to the report/UI, as they can contain duplicate users and it won't match up.
- When you compare to the Export Report you are checking the right section - for example, if you are querying the last 7 days, make sure to check the "Weekly" value for the last day of the results as that will have the correct calculation.
- If not splitting by date, make sure you are querying by 1-day, 7-day, or 28-day date ranges to avoid errors (especially in Data Studio).
If you checked these and are still seeing discrepancies, see the section Escalating to Support below for what to do next.
Data Studio Discrepancies with Filters
Google fixed this at the end of July 2019, so we are now passing the filter as part of the request and you should be able to get accurate reach values. This does expose a different issue, that certain filter can now lead to this error (which Data Studio usually shows as a "No Data" or "null" value):
So you either need to be very specific with the filter so that it only finds one item or you add additional dimensions to make it more unique.
Grand totals are calculated from Page totals
Summary row as in selection "Show summary row" (= Grand totals) may seem off, for example when you have the query on Post level metrics. Summary row results are actually a whole different query, showing Page totals. This is intended behavior, but may look like a discrepancy.
Escalating to Support
If you've checked everything above and you still believe that Supermetrics has a discrepancy with the results in Facebook Insights UI, then please gather the following information and place it in a new Support request:
- A screenshot showing the incorrect value in the Supermetrics product
- A screenshot showing the configuration of the query in Supermetrics that is incorrect (or the query ID if using Google Sheets)
- A copy of the Export Report .XLS file for the same date range as the Supermetrics Query. You can get this from the Insights UI here:
Then set the date to match your query and select if Page or Post data, depending on what kind of information you are getting.