Google Analytics has different methods for counting and delivering your data to us. Due to this, discrepancies can occur between different representations of that data in our platform, and also when comparing the data in our platform with Google Analytics directly. 

All of the numbers we display come directly from the Google Analytics API. The discrepancies relate to the treatment of duplicate website visitors: whether they're actually counted, and whether those duplicates are first split into different categories (and then tallied) or simply reported as a total. 

In this article, I'll refer specifically to the "users" metric, but the explanation also applies to "session" and other metrics where duplicate visitors are possible.

As an example, take a look at the image below. The number in the pie chart is being sent to us directly from Google Analytics via their API. The number in the box below the chart is calculated by us; we simply add up the total number of users for each channel, as sent to us via the Google Analytics API. Both of these numbers can and will differ from the "user" metric that you'll find by looking in your native Google Analytics account, which we discuss in a separate article.

To explain further, that number in the pie chart includes only unique users, but may include duplicates. Google defines those terms in different ways, and in the native GA interface, the number shown includes only unique users, and NOT duplicates.

As for the number shown in the box below, Google sends us data (again, via their API) on the "user" metric broken out by Channel. Similar to the number in the pie chart, those per-Channel numbers include unique users only, and may include duplicates. But here's where the difference comes in: For the per-Channel numbers that Google's API sends us, there can be additional duplicates because some users arrive by more than one different channel during the time period.

So when we add all of the numbers for every channel together to come up with that total "users" metric in the box, not only are duplicate visits possibly being recorded; duplicate unique users among different channels are being recorded as well.

Tl;dr version:

  • Number shown in GA: Unique visits, duplicates removed
  • Number shown in graph: Number from API, may include some duplicates
  • Number shown in box: Number from API, may include duplicates, may also include additional duplicates due to users who visit via different channels

Along with the "users" metric, this can also be the case for sessions, pageviews, and other metrics related to user channels and engagement.

Why do we report data in multiple ways?

Because this type of data is not black and white, and can appear to report different things depending on how it's calculated. By reporting the data using multiple calculation methods, we're giving you a more accurate picture of what's truly happening with the campaign in question. With a more accurate picture, you can decide for yourself which data set most accurately represents the true state of the site's traffic.

Did this answer your question?