Mapping the website and mobile app audiences of Russia’s foreign communication outlets, RT and Sputnik, across 21 countries

---

Implications

Following Russia’s invasion of Ukraine on 24 February 2022, policymakers worldwide have taken measures to curb the influence of Russia’s foreign communication outlets, RT and Sputnik. Television companies in Australia (Meade, 2022) and the United States (Darcy, 2022), as well as the Canadian government (Government of Canada, 2022), banned RT’s TV broadcasting for its pro-Russian coverage of the war in Ukraine (Darcy, 2022; Meade, 2022) and, in the case of Canada, noncompliance with national broadcasting standards (Government of Canada, 2022). The Council of the European Union (EU) (Council of the EU, 2022; European Union, 2022) and the UK (GOV.UK, 2022) went one step further and banned the dissemination of RT/Sputnik’s content via any means, including the outlets’ official websites and mobile apps, though extant research suggests that niche news media reach small audiences compared to mainstream media via websites/apps in Europe and the United States (see Altay et al., 2022; Nelson & Taneja, 2018). EU and UK policymakers justified their decision by claiming that RT/Sputnik are part of Russia’s disinformation campaign and have been essential to the preparation and support of Russia’s invasion of Ukraine (European Union, 2022; GOV.UK, 2022). 

The content of RT and Sputnik has been extensively researched (e.g., Deverell et al., 2021; Kragh & Åsberg, 2017; Ramsay & Robertshaw, 2019; Yablokov, 2015). For example, even before Russia’s invasion, scholars found that RT had published conspiracy theories (Yablokov, 2015) and mis- and disinformation (Elswah & Howard, 2020) alongside a range of other types of content. 

We know less about their social networking sites’ (SNSs) audiences (Crilley et al., 2020; Orttung & Nelson, 2019) and virtually nothing about their website/app audiences. Relevant academic studies that have investigated the outlets’ multiplatform audiences (including SNSs and websites/apps) have drawn on either cross-national (see Newman et al., 2020) or national surveys (see Müller & Schulz, 2021; Wagnsson, 2022). 

These survey-based studies, in particular the cross-national ones, provide only a coarse-grained picture of the popularity of these outlets and have at least three disadvantages. The first disadvantage, as extant research has shown (e.g., Guess, 2015), is that respondents typically struggle to accurately recall their online media consumption patterns (recall bias). Secondly, survey participants may give answers that correspond with social norms (social desirability bias; see Lavrakas, 2008). RT and Sputnik have been criticized for spreading mis- and disinformation (e.g., Elliott, 2019). Therefore, it appears plausible that estimates of RT/Sputnik’s audiences based on surveys could be compromised by social desirability bias and recall effects. Thirdly, these survey-based studies did not capture audience reach and demographics by distribution channel. Consequently, we know very little about how their websites/apps contribute to the visibility of the two outlets. 

RT has traditionally focused on SNSs to disseminate its content, particularly the video-sharing platform YouTube (Orttung & Nelson, 2019). Orttung and Nelson (2019) report that, as of 2017, RT’s English-language channel performed on YouTube slightly better than even the leading global news channels, CNN and the BBC. Nonetheless, particularly after YouTube’s decision to deplatform the two outlets in March 2022 (Reuters, 2022), RT/Sputnik’s websites/apps can be considered central to their informational ecosystems because a large proportion of their SNS posts link back to their websites/apps. 

There is currently no academic publication that has scrutinized the website/app audiences of RT/Sputnik. Moreover, we know of no cross-nationally comparative research on the demographics of the outlets’ audiences (for survey-based national investigations, see Müller & Schulz, 2021; Wagnsson, 2022). Finally, the development of their audiences over time, to our knowledge, has never been investigated in peer-reviewed academic literature. To help fill these gaps, we investigated the reach of RT/Sputnik’s websites/apps; channels that, according to EU and UK policymakers, supported the invasion of Ukraine (European Union, 2022; GOV.UK, 2022). In particular, we analyzed the outlets’ website/app audiences using Comscore data from 21 countries, focusing on (1) audience size, (2) reach, (3) demographics, and (4) changes over time between 2019 and 2021.

Our findings show that, in the quarter before the invasion, RT/Sputnik did not have monthly reaches of more than 5% of the digital populations of any of the 21 countries via websites/apps. Between 2019 and 2021, averaged across these countries, the outlets’ website/app reach remained approximately constant. However, there were some short-lived peaks of over 5% in some countries. RT/Sputnik’s website/app reach and audience size differed substantially between countries. Across countries for which data was available, RT’s average website/app reach (1.05%) was larger than Sputnik’s (0.79%). Moreover, both outlets reached more men than women (RT: 1.40% male vs. 0.73% female; Sputnik: 0.98% male vs. 0.66% female), and their website/app reach increased with audience age (see Figures 4 and 5). 

Our findings have at least four implications. Firstly, by comparison with their alleged popularity on SNSs (see Orttung & Nelson, 2019), we demonstrate that the two outlets each attracted small audiences via their websites/apps immediately prior to Russia’s invasion (in all countries, less than 5% of the digital population monthly). In contrast, taking two UK news brands as examples, BBC News Online and MailOnline each have an average monthly website/app reach of more than 50% of the digital population in the UK (Thurman et al., 2021). This finding is in line with previous research that suggests that niche and untrustworthy news media reach small audiences compared to mainstream media via websites/apps in Europe and the United States (see Altay et al., 2022; Nelson & Taneja, 2018). Comparing the website/app reach of foreign news outlets originating in an authoritarian country, such as Russia, with the website/app reach of domestic news outlets in a democratic country, such as the UK, has limitations, but it does help show the relative size of RT/Sputnik’s audiences. We recommend future research compare RT/Sputnik with the foreign communication outlets of other authoritarian states (e.g., China, Iran, Qatar). Despite finding rather small website/app audiences in our sample of countries, RT/Sputnik’s impact on Western audiences and politics should, however, not be underestimated. The outlets’ websites/apps may be less important than their SNSs as vehicles to reach audiences (Orttung & Nelson, 2019), and RT/Sputnik may fulfill a far broader range of tasks for Russia’s authoritarian elites.

Secondly, adding to previous findings on RT/Sputnik audiences (e.g., Crilley et al., 2020; Wagnsson, 2022), this study finds that the outlets’ content was more likely to be consumed by men and that their website/app reach increased substantially in older age groups. Future research could investigate why it is that these groups follow Russian state media more than others.

Thirdly, as our findings show, between 2019 and 2021, RT/Sputnik’s website/app audience size and development varied substantially across the 21 countries. Accordingly, we argue that the impact of the two outlets via their websites/apps has to be assessed for each national context separately. In some countries in the EU, such as Ireland, Finland, and Italy, both RT and Sputnik’s monthly website/app reach was, on average, less than 1% prior to Russia’s invasion. In other countries, such as Spain (RT: 3.94%; Sputnik: 2.63%), Germany (RT: 3.18%), and France (RT: 2.32%), it reached higher proportions of the populations. 

Finally, we encourage future research to investigate audience behavior on RT/Sputnik’s websites/apps more deeply. For example, it would be interesting to compare, in Western democracies, the behavior of the website/app audiences of the two outlets with that of alternative, quality, and other foreign communication outlets from authoritarian countries. Such research could advance extant literature on fake and alternative news outlets (see Altay et al., 2022; Müller & Schulz, 2021; Nelson & Taneja, 2018; Schwarzenegger, 2021, 2022; Taneja, 2020) and inform debates on the consequences of RT/Sputnik’s presence in democratic public spheres. 

Findings 

Finding 1: RT/Sputnik’s websites/apps’ audience size and reach differed substantially across 21 countries and between the two news outlets; however, neither outlet, at least via their websites/apps immediately prior to Russia’s invasion of Ukraine, reached monthly more than 5% of the digital populations of any of the 21 countries.

To understand how large RT’s and Sputnik’s website/app audiences were prior to Russia’s invasion (RQ1), we averaged monthly audience data for the last quarter of 2021. As Figure 1 illustrates, the average number of monthly unique visitors (MUVs) to RT/Sputnik’s websites/apps varied substantially across the investigated countries.¹Average monthly unique visitors in Hong Kong are too few to appear in Figure 1. Similarly, RT/Sputnik’s monthly reach varied greatly across countries but was below 5% in all countries (see Figure 2).²The reach of each outlet was also below 5% for the averaged monthly data from March 2019 to December 2021 inclusive (although reach occasionally exceeded 5% in some country months), and if missing values were replaced with zeros or with half the lowest recorded reach figure for that country. The average website/app reach across all countries for which data was available (RT: n = 19; Sputnik: n = 10) was larger for RT (1.05%) than for Sputnik (0.79%).³The reach of RT was also larger than Sputnik for the averaged monthly data from March 2019 to December 2021 inclusive (0.99% vs. 0.78%) and if missing values were replaced with zeros (0.95% vs. 0.37%) or with half the lowest recorded reach figure for that country (0.95% vs. 0.37%). As Figure 2 shows, RT’s audience reach was commonly larger than Sputnik’s in countries where both outlets had an audience of sufficient size to be measured by Comscore (see Methods section), except in Argentina and Italy.

Finding 2: RT/Sputnik’s websites/apps were more likely to reach men and those in the older age groups.

To examine how the reach of RT/Sputnik’s websites/apps differed by gender and age prior to Russia’s invasion (RQ2), we again averaged audience data for the last quarter of 2021. As Figure 3 illustrates, across all countries for which Comscore data was available, both RT’s and Sputnik’s websites/apps had a higher average reach among men (RT: 1.40% male vs. 0.73% female; Sputnik: 0.98% male vs. 0.66% female).⁶The reach of both outlets was also higher among men for the averaged monthly data from March 2019 to December 2021 inclusive (RT: 1.30% vs. 0.71%; Sputnik: 1.03% vs. 0.58%), and if missing values were replaced with zeros (RT: 1.26% vs.0.66%; Sputnik: 0.46% vs. 0.31%) or with half the lowest recorded reach figure for that country (RT: 1.26% vs. 0.66%; Sputnik: 0.46% vs. 0.31%). There was no country in which RT/Sputnik’s website/app reach was larger among women than men, though the differences were sometimes very small.

The age groups provided by Comscore for RT and Sputnik vary by country; hence, we present data on each country for which Comscore data was available separately. As Figures 4 and 5 illustrate, although the differences were sometimes too small to be meaningful, in most cases, the average reach among the oldest age group was larger than among the youngest. The exceptions were RT in India and Sputnik in Colombia and Japan, but again, the differences were negligible.

Finding 3: The reach of RT/Sputnik’s websites/apps varied substantially between countries from March 2019 to December 2021. However, across all countries, the audience reach of both outlets remained approximately constant.

To study how the global audiences of RT/Sputnik’s websites/apps changed between 2019 and 2021 (RQ3), we averaged percent reach data across all 21 countries by news outlet. Comparing the earliest equivalent quarter (time period 1 (t1): October–December 2019) with the last (time period 2 (t2): October–December 2021), we found that although the audience reach of both outlets varied substantially in individual countries, it remained approximately constant for RT and decreased slightly for Sputnik when all 21 countries included in our sample were considered in aggregate (percentage point difference between first and last quarter: −0.06% for RT and −1.14% for Sputnik).⁷The reach of RT is also approximately the same at t2 compared to t1 if missing values are replaced with zeros (−0.06 % points) or with half the lowest recorded reach figure for that country (−0.06 % points). The reach of Sputnik at t2 compared to t1 decreases further when we replace missing values with zeros (−0.76 % points) and with half the lowest recorded reach figure for that country (−0.76 % points). 

However, the outlets sometimes had a much larger reach than the averages suggest. Sputnik’s website/app reach in France was almost 15% in early 2020, and RT’s website/app reach in Argentina peaked at around 6% in March 2020.⁸These peaks seem to coincide with major news events, such as the beginning of the Covid-19 pandemic. Hence, we should avoid over-interpreting the downward trend we see in some countries because it could quickly disappear if coverage of a specific story gains audience traction (assuming access is not blocked).

Figures 6 and 7, which plot the monthly data over time for each outlet, show how the size of the changes differed by country. RT’s audience reach increased most in Spain (+1.88 percentage points) and Germany (+1.63 percentage points) and decreased most in Mexico (−1.85 percentage points) and Ireland (−0.86 percentage points). Sputnik audiences increased most in Spain (+1.69 percentage points) and Argentina (+0.56 percentage points) and decreased most in France (−9.87 percentage points) and Japan (−0.90 percentage points).

Methods 

To answer our research questions (see beginning of the article), we used data from Comscore, an analytics company that provides audience estimations of website/app entities by combining data from samples of online-recruited panelists and server-side measurements. The company collects internet activity, including “private browsing,” of desktop, tablet, and mobile panelists automatically and weights the data to reduce demographic/behavioral bias and make the sample more representative of a country’s internet users (Comscore, 2021). We accessed the MMX Multi-Platform product, which has been used for audience research in the past (e.g., Thurman et al., 2022). Comscore data has also been used as a primary source in other academic fields, including marketing (e.g., Chesnes et al., 2017) and medical informatics (e.g., Kim et al., 2016). In each country, we analyzed data from the Russia Today and Sputnik Sites entities, which combine data on all relevant domains, apps, and language editions (for further details on the retrieved data, see Appendix). 

Studying the website/app audiences of Russia’s main foreign communication outlets using Comscore data has several advantages. RT/Sputnik’s websites/apps are central nodes of Russia’s foreign content dissemination strategy because all RT/Sputnik’s SNS accounts link to them. Because SNSs can rise and fall in popularity, the usage of websites/apps may be a more consistent measure of media outlets’ popularity over time. In addition, RT/Sputnik’s websites/apps contain the outlets’ core content, whereas their SNSs mostly carry snippets of it. Finally, although RT/Sputnik’s SNS audiences can be gauged by public follower counts and engagement metrics,⁹Although RT has been accused of inflating these metrics in the past (U.S. Department of State, 2022). data about their website/app audiences is not freely available. To our knowledge, there has been no academic study that has scrutinized RT/Sputnik’s website/app audiences.

Because we purchased access to a subset of Comscore’s data, country selection was somewhat limited. For example, we did not include any Arabic-speaking countries in this study. In the following section, we provide details on the measurements used in this study and how these were analyzed to answer our research questions.

For RQ1 and RQ3, we used data on monthly unique visitors (MUVs) and percent reach, which was provided for each news outlet, country, and month. MUVs are the estimated number of unique individuals that visited the entities’ websites and apps in a month (Comscore, 2021, p. 90). The percent reach estimate is defined as the “percent of total universe accounted for by total site visitors” (Comscore, 2021, p. 91) in any given month.¹⁰In other words, the measure is Comscore’s estimated digital population of a country divided by their estimate of an entity’s MUVs. We limited our analysis to adult users, which we defined as all those aged 18 and over. However, in some countries (e.g., Italy), the lower bound of the youngest adult age group in the data starts at 15 (see Figures 4 and 5).¹¹In Argentina, Chile, Colombia, and Ireland, data for the relevant “Persons: 15+” or “Persons: 18+” category was not available, so data is for “Total Audience.” To investigate how large RT/Sputnik’s website/app audiences across the countries under investigation were (RQ1), we calculated the average number of MUVs and percent reach by country and news outlet in the last quarter before Russia’s invasion. For all RQs, if Comscore’s minimum reporting standards (MRS) were not met, the MUVs and percent reach were reported as N/A and, hence, not included in the average for a country/news outlet in this study. However, we also re-ran the analysis with imputed values to check the robustness of our findings (see last two paragraphs in this section).

In order to analyze how RT/Sputnik’s website/app audiences developed between 2019 and 2021 across the 21 countries (RQ3), we focused on percent reach. The percent reach estimate is more valid than the MUV estimate in this context because the number of total internet users in a country might change over time. Drawing on the difference between the averages for October–December 2019 and October–December 2021, we identified the countries where RT/Sputnik gained/lost in reach (defined as a positive/negative percentage point difference between the first and last quarter). 

To answer RQ2, we relied on Comscore’s demographic data. In particular, we used the percent reach metric, which was provided by news outlet, country, month, and demographic groups of gender (male/female) and age (age groups differed by country). 

To understand how the website/app audiences of RT/Sputnik compared in terms of gender in all 21 countries, we used data on all adult male and female visitors.¹²In Ireland, data for the relevant “Males: 15+” and “Females: 15+” category was not available, so data is for “All Males” and “All Females.” First, we calculated the average % reach per group (female, male) in a country for the last quarter of 2021. Thereafter, we calculated the average across all countries per group.

For age, Comscore provided data with percent reach per age group (for example, 35–44 years or 35+ years), country, and month. The age groups provided by Comscore differ by country. Where available, we analyzed data from all non-overlapping age groups that include those aged 18 and older (see Figures 4 and 5). Because age groups differ by country, it was not possible to analyze the pooled data. Instead, we studied the data by country, comparing the percent reach for the youngest age group to the oldest.

In the main analysis, if Comscore’s MRS were not met, we treated values as missing in line with how Comscore reports the data. However, it is very likely that values were missing because they were too small. Hence, we also re-analyzed the data with missing values recoded as zeros and as half the lowest reported value for that country from March 2019 to December 2021 inclusive. This imputation allowed us to include all 21 countries in the analysis (instead of RT: n = 19; Sputnik: n = 10) and arguably better reflects the underlying reality. We also re-analyzed using data from March 2019–December 2021 inclusive, instead of just the last quarter before the invasion. These results are footnoted in the Findings section and do not substantively change our interpretations. 

It is important to recognize that the differences we describe in the Findings section are very small in percentage point terms and should be interpreted with caution. In large part, differences are small because of the small overall reach of RT/Sputnik. Because Comscore’s estimates are based on data from a sample of online-recruited panelists and server-side measurements, it was not possible to estimate error in the standard way. However, because Comscore’s panels are typically very large (e.g., approximately 300,000 in the United States) and because server-side measurement does not rely on sampling, there is good reason to expect sampling error to be much lower than that typically associated with surveys. As a broad, tentative rule of thumb, we do not consider differences of less than ±0.25 percentage points to be meaningful. However, let us reiterate again that because it is not possible to estimate error, small percentage point differences should be interpreted with caution.