The study of 315,000 VK profiles for NATO StratCom

How we investigated the effects of banning VKontakte in Ukraine

About the client and the research goals

NATO Strategic Communications Centre of Excellence (NATO StratCom COE) is a NATO-accredited international military organization. It was founded in 2014 and consists of seven member states – Estonia, Germany, Italy, Latvia, Lithuania, Poland, and the United Kingdom.
The organization’s mission is to improve strategic communications capabilities within the Alliance and Allied nations.

In June 2018, NATO StratCom turned to our company for assistance with new research concerning security challenges in the online environment and the role of governments in countering them.

The object of our study

In May 2017, the largest Russian social network, VKontakte (VK), was banned in Ukraine to protect the national security of Ukraine from pro-Russian propaganda. After that, according to Google Trends, there was decrease in VK users of almost 80%.

 

But the effects of the VK ban had to be studied from other different perspectives to understand the effectiveness of such a government reaction to propaganda.

We designed a study of 315,697 Ukrainian active VK user profiles of different age groups and locations that took place from 1 May 2016 to 14 June 2018. As some territories are occupied by Russia, we also studied two regions separately: the government-controlled areas (GCA) where the VK ban was instituted, and the non-government-controlled areas (NGCA), where the ban was not imposed.

Our specific goal was to analyze changes in posting dynamics, user demographics, and the ideological topics used in posts before and after the ban.

Instruments used

Social Media Intelligence tools, neural networks, clustering, social graphs, user behavior analysis.

Facts we discovered

1. Posting dynamics

To study this aspect, we developed proprietary Python scripts for information retrieval, which allowed us to investigate almost 20 million posts.

As a result, we managed to determine three waves of posting dynamics in VK during the study period.

The first wave was before the ban, when the total number of posts reached 101,000 per day. The second wave was after the ban, when the number decreased by 53% to 38,000 on average. Since April 2018, when the third wave started, traffic has decreased by 10,000 and stabilized at that level.

2. User behavior analysis

With the help of social graphs, we managed to conduct user behavior analysis and investigated:

  • who left the VK after the ban
  • who remained on the social network
  • who was most influential, and how their connections changed.

 

 

The research results showed that, after the ban, VK users became more connected. Users from the GCA (government-controlled areas) had 16% more friends on average after the ban than they had before it.

We also found that users from the GCA who continued to use VK after the ban were 4.2 years younger than users from the non-government-controlled areas (27 vs. 31 years old).

3. Topic analysis of posts

During the research, we used a clustering algorithm that could identify account postings on ideological issues. We found that the number of ideological posts after the ban increased by 1.22 times compared to the period before the ban.  

 

According to the results of our clustering process, a new topic, ‘Russian news’, appeared. ‘Pro-Russian propaganda’ notably increased, while the share of ‘Ukrainian news’ decreased. These changes can be explained by moving to the pro-Russian infosphere.

Ideological users became significantly more connected – after the ban, the average number of friends for an ideological user grew from 197 to 501.

The results we obtained

A team of 5 people worked on this research for a month.

The VK network’s popularity decreased markedly in Ukraine. But besides this predictable fact, many other aspects were discovered:

  • After studying 20 million posts, we determined three waves of posting dynamics and noticed a new tendency for user exodus after April 2018.
  • After comparing the dissemination of VK in different territories, we concluded that it is now only 19% lower in the government-controlled areas than in the territories occupied by Russia.
  • With the help of social graphs, it was determined that, after the ban, VK users became more connected on average and began consuming more information from a greater number of groups.
  • The VK audience became younger, and its average age differs in GC and NGC territories.
  • Using a clustering algorithm, we could identify ideological users who continued to use VK. They became more active in posting compared to the average user. The share of ideological posts increased – from 36% before the ban, to 52% after the ban.

The research became a part of a NATO StratCom study devoted to the role of governments in countering critical security challenges. It was included in the publication “The Effects of Banning the Social Network VK in Ukraine”.