There are 3 different ways that the views on blog posts can be found. There is a counter on the Blog post itself, Platform Insights and the R603 report. These numbers can have differences due to the way each are calculated. This document aims to explain the reason by any differences to ensure accurate reporting on the engagement of Blog posts.
Blog Counters
On a blog post itself, you can see a count of views.
This count will not represent the exact number of unique views, but rather to give you a real-time trend. If you have a blog post which has 100 unique views, in reality, this blog post may have between 95-105 views.
This counter is implemented using a HLL algorithm, The idea of the HLL algorithm is to offer a solution to the count-distinct problem often found in all modern web applications. More for the count-distinct problem can be found here.
HLL makes it easy to estimate the number of unique values within a large set of elements - known as the cardinality of a set. Look at the following set of numbers {1, 2, 3, 3, 1, 6, 7, 6, 9, 1}. The cardinality of it is 6 because the unique elements are {1, 2, 3, 6, 7, 9}.
Therefore, this counter is to be used as a rough trend of how many views the blog post has but it may not be 100% accurate.
Platform Insights
Platform Insights is a more reliable way to understand the amount of views on a blog post but does have some considerations.
There may be a scenario where there is a difference in the 'Total Views' section and the number of views counted in the 'Most Popular Charts' section.
The Most popular posts chart retrieves the blog posts published in the selected period (set in the filter at the top of the page), while the Total views on posts and Unique views on posts charts count total and unique views for that period, and it doesn’t matter whether the posts are published within the selected period or not.
Let's look at an example:
Platform Insights was ran for 1st April-1st June.
If you sum all total views in the Most Popular Posts, you will get 612 (451 + 91 + 70 = 612), which is with 6 less than Total views on posts chart - 618
In this example, this is because one of the blog posts which had 6 views was published outside of the timeframe that the Platform Insights was run for (before 1st April).
One other important thing to note is that the data in Platform Insights is not real-time. It is updated daily. In practice, this means that the same blog post with the same 100 unique views may have received 10 more unique views today and now it says 110, but Platform Insights data will still show 97, or 105 or 102 unique views until the data source is updated, at which point it will reflect the correct number of views.
R603 - Blog Articles Views Report
This report once again gives an accurate amount of views on a report. However it's key to note that this report can not always be comparable to Platform Insights report for counting the number of Unique Views. Here is why:
When running the R603 Report for a longer period (EG 2 months - January & February 2025) - the report that is generated displays ALL views that took place over that period, irrespective of when the Blog Post was published.
In the report above, you can see blog posts from 2024 in there, even though the report was ran for January 2025 - February 2025; This is because there were still VIEWS on that blog post between Jan-Feb 2025, even though the blog post was published before.
Secondly, the views (counted on the right hand side under Jan 25, Feb 25) are unique views for each month during that period.
That means that if USER A read the blog Post in January, that would count as 1.
If the same user read the same blog post in February, that would be a Unique view for February and does not affect January's count, resulting in 2.
The logic is not the same on Platform Insights -- the unique view on a blog post by the same user would only count as 1, as the period is January to February, and we would only count a unique view once for the entire period. Therefore they cannot be compared, but both offer valuable insights, just different.