Author: Mika Desblancs-Patel

Organization: Media Ecosystem Observatory

Key takeaways:

Following the indictment on Sept 4th, we identified a surge in hashtags like #PierrePutin claiming that Poilievre is acting in the interest of the Russians.
The surge of PierrePutin hashtags lasted for 12 days hitting a peak of 1475 posts which accumulated 1,303,303 views.
We identified a set of influential accounts sharing these hashtags, yet did not find sufficient evidence to suggest this narrative was manufactured by a subset of accounts or by bots.

Introduction

When the news broke that Tenet Media and six right-wing influencers had received money from Russia, many worried about how successfully the scheme had spread misinformation, polarized Canadians or eroded trust in their elected officials and large media companies. Beyond the event , there is growing concern that the way these revelations are handled can be substantively damaging, perhaps even more so than the campaign itself.

In a previous update, a survey analysis of awareness and importance of Tenet Media influencers in Canada, we report that one in four Canadians are aware of the case. In this update, we try to uncover how Canadians handled news of the incident by looking at how they engaged with the story on social media, specifically X/Twitter (referred to as X for the remainder of the post). We focus on X mainly because this story was most widely known among X users (38%).

We investigate how Canadians talked about the Tenet Media case on X by exploring the following questions:

How much were Canadians talking about the incident on X?
What types of discussions emerged about the story? Which hashtags were used?
How did major hashtags about the story emerge?
Was the emergence of #PierrePutin manufactured?

How much were Canadians talking about the incident on X?

To address this question, we assessed the presence of “Tenet Media” in conversations on X. To do so, we collected posts using X’s advanced search mentioning the terms “Tenet Media” or in discussions mentioning “Tenet Media” 14 days before and after Sept. 5th 2024 (August 22nd to September 19th 2024). In total, we collected 16,563 posts with Figure 1 showing count per day. Among all posts, 5,888 (~36%) had poster location data available, of which 1,330 (~23% of all posts posted by users whose location was public) were from users in Canada. Unsurprisingly, we find that the majority of the discussion on X happened immediately after the indictment was released publicly.

Figure 1: Number of posts per day on X mentioning “Tenet Media” for all users on X as well as users whose public location is in Canada. The X-axis shows the days and the Y-axis the post count. The numbers above the blue and orange bars are the post counts for that day for all posts and those posted by users whose public location is in Canada.

To evaluate the amplitude of this story in comparison to other stories that emerged on the same day, we contrasted engagement with the NDP announcement on Sept. 4th 2024 of the end the Supply and Confidence agreement with the Liberals. We chose this story because a) both the DOJ’s and the NDP’s announcement happened on the same day, b) the story was heavily discussed across outlets in Canada and we knew it was popular. We applied the same collection method as in the previous section and looked up the presence of “supply and confidence” in conversations on X over the same period. In total, we found 977 posts from users in Canada. We find that the Tenet Media story was more widely talked about on X than the supply and confidence agreement.

To further evaluate differences in engagement between the two stories, we looked at the discussion using the Canadian Digital Media Research Network’s list of influential accounts in Canada. Because the amount of posts about each topic is much smaller and manageable, we looked up a wider range of keywords associated with each story to capture as many as possible.

Figure 2: Distribution of the posts about Tenet Media versus the end of the ‘supply and confidence’ agreement from influential Canadian X accounts by post count (number of daily X posts) and total engagement (likes + quotes + reposts + replies + views)

Looking at Figure 2 we find that, among accounts in our list, the ‘supply and confidence’ story generated more acute engagement than the Tenet Media story during the period of September 4-19, 2024. At its peak on Sept. 4th, the ‘supply and confidence’ story generated 97 posts (cumulative total engagement 1,015,755) while the Tenet Media story, on Sept. 5th, received almost half, 51 total posts (cumulative total engagement 454,737). The post count for the supply and confidence agreement, however, rises and falls quickly (219 total posts for the period) while the Tenet Media story was more sustained (196 total posts for the period). As demonstrated in Table 1, the supply-and-confidence received more overall attention as it was posted about 219 times and saw a total engagement of 1,934,009. Meanwhile, the Tenet Media story was posted about 196 times with a total engagement of 1,662,230.

Table 1: Post count, number of unique users, and total engagement per account type for both the Tenet Media story (“Tenet media") and the supply-and-confidence story (Supply & conf) for the period of Sept. 4th till Sept. 19th, 2024.

Table 1 also demonstrates the differences in the types of accounts spreading the story. The ‘supply and confidence’ story was primarily, though not overwhelmingly, spread by Canadian news outlets, sharing 56% of the posts (123 posts out of 219) and capturing 52% of all engagement (1,007,482 out of 1,934,009). On the other hand, the Tenet Media story was shared overwhelmingly by Canadian influencers, making up 85% of the posts (167 posts out of 196) and 67% of total engagement (1,109,489 out of 1,662,230).

We find that overall the supply and confidence story was more popular than the Tenet Media story among accounts in our list of influential accounts. This divergence in attention paid to the Tenet Media story captures a difference between influencer-driven social media (who were much more interested in the influencer-focused Tenet Media story and can explain why general users followed suit) and more traditional media and political actors (who were more interested in the institutional politics supply-and-confidence story).

What types of discussions emerged about the story?

To better understand discussion associated with the story, we identified popular Canada-focused hashtags in the dataset of posts we collected. We evaluate these hashtags to focus on how the conversation matters in Canada, identify posts associated with the broader conversation, gauge opinions and identify dominant narratives on a topic.

The most popular hashtags were stance-neutral (e.g. #TenetMedia, #Russia, #cdnpoli), while many expressed a stance (e.g. #PierrePutin, #IStandWithTrudeau). We cataloged all popular hashtags which expressed a Canada-specific stance and came up with a list of hashtags to follow over the same period. We then generated a new dataset by collecting all posts referencing one or more of these hashtags in the same date range (these posts did not necessarily contain ‘Tenet Media’). In total we collected 46,178 unique posts, see the methodology for the full list of hashtags.

Image 1: Example of a post about the Tenet Media story posted by a Canadian user expressing a politically charged hashtag

Among this new dataset of posts, we broadly grouped the hashtags into in 10 mutually exclusive categories:

Pro-Trudeau (e.g. #IStandWithTrudeau2025, #CanadaStandsWithTrudeau)
Anti-Poilievre (e.g. #SickOfPoilievre, #PierrePoilievreIsUnelectable)
Anti-conservative Party (e.g. #NeverVoteConservative, #NeverCPC)
PierrePutin i.e. posts linking Poilievre to the Russian propaganda campaign (e.g. #PierrePutin, #PierrePoilievreIsARussianAsset)
Anti-“Freedom Convoy” (e.g. #FreeDumbConvoy)
Anti-NDP (e.g. #SelloutSingth)
Anti-Doug Ford (e.g. #DougFordIsCorrupt)
Anti-Danielle Smith (e.g. #DanielleSmithIsUnfitToLead)
Media collusion (e.g. #CdnMediaColludes)
More neutral hashtags (e.g. #cdnpoli, #canada, #RussianInterference)

Figure 3 shows the counts of posts containing a hashtag in each of the categories (excluding neutral hashtags) during the period before and after the announcement of the DOJ indictment.

Figure 3: Daily count of tweets containing a hashtag in each of the categories (excluding neutral hashtags). The color code is on the right in the “Hashtag Categories” section

Of the 10 groups of posts, we find that two exhibit significant changes in evolution trends between the period before and after the DOJ indictment came out: the pro-Trudeau, and PierrePutin hashtags. Usage of both these hashtags changed on or around September 4th.

We next try to understand whether the increased use of these two hashtags associated is because of the Tenet Media story or another event in the information ecosystem?

The rise of pro-Trudeau hashtags appears to coincide with the NDP’s announcement ending the Supply and Confidence agreement with the Liberals on September 4th. To disentangle the two, we assessed the content of posts containing pro-Trudeau hashtags on Sept. 5th 2024, the day uses of the hashtags spiked. We found that approximately 14% of the 1281 posts contained mentions of ”NDP,” ”Jagmeet,” ”Singh,” “supply and confidence” or “agreement.” On the other hand, only ~1.8% of the same 1281 posts contained words related to the Tenet Media story like “tenet,” “Russia,” “Putin,” the “DOJ,” “influencer,” “fund” and “disinfo.” This spike appears to be much more strongly associated with the end of the NDP agreement than the Tenet Media story.

The rise of the PierrePutin hashtags, on the other hand, seem much more likely to be associated with the news of the Russian disinformation campaign. First, we found 28% of posts containing PierrePutin hashtags had words related to the story (we excluded the word “Putin” in our keyword search to avoid catching the “Putin” in e.g. #PierrePutin) on the 11th of September, the day of the highest point in PierrePutin hashtag use. Furthermore, the second most used hashtag in this group is #PierrePoilievreIsARussianAsset and is clearly about the story. Finally, hashtags in this category went from virtually unused before Sept. 4th to trending on X a couple days after. Among posts collected in our dataset, before Sept. 4th (from Aug. 22 to Sept. 4th), we find only 11 posts by 5 users containing hashtags in the PierrePutin hashtag while, at its highest point on Sept. 11th, there were 1475 posts by 519 users. The combination of the presence of related keywords in posts containing the hashtags, thematic proximity between the Tenet Media story and the hashtags’ content, along with the sharp increase in use is strong indication that the Tenet Media story prompted increased use of the hashtag. From the strong increase in PierrePutin hashtags in the days following the release of the indictment, many posts on X argued that Poilievre was acting in the interest of the Russian government.

How did PierrePutin hashtags emerge?

We next evaluate the extent to which the use of hashtags associating Pierre Poilievre and Russian information operations was produced organically and by Canadian voices. We find sporadic use of #PierrePutin prior to the Tenet Media Incident, with 11 posts from 5 users containing hashtags in the PierrePutin category from August 22 to September 4. Six of these posts argued it was hypocritical for Poilievre to celebrate Ukrainian Independence day while not supporting Canada’s aid to Ukraine. Going further back, we find that the first instances of the PierrePutin hashtag happened as early as March 2014 (see Image 2), but again that hashtag were not used at any meaningful scale.

Image 2: The first two tweets on Twitter using the #PierrePutin hashtag

Evaluating the period following the Tenet Media incident, Figure 4 shows a full evolution of the hashtag in our date range by indicating the daily count of posts containing the #PierrePutin and associated hashtags within the PierrePutin category.

Figure 4: Total number of tweets per day containing PierrePutin hashtags (the numbers above the bars indicate the count of tweets for that day)

As shown, we find hashtags in this category started to rise on the 4th of Sept. with 13 posts, then 30 the next day, 304 the following, before climbing to 747 posts on the 7th and continuing its ascension to 1,475 on the 11th.

Figure 5 captures the evolution of the number of views of posts with these hashtags. First the total number of views of posts containing these hashtags spikes on the 8th of Sept, three days before the spike in the number of posts on the 11th. On the 8th, the hashtag #PierrePutin was so popular it was trending on Canadian X. Furthermore, the evolution of daily total views drops off faster than the daily post count with total views dropping to under half of its peak the very next day, and only going down afterwards.

Figure 5: Sum of the total number of views per day of tweets containing PierrePutin hashtags (the numbers above the bars indicate the count of tweets for that day)

So who was using the PierrePutin hashtags? When were they using it and how?

We identified some trending posts which drove attention to the narrative. The first one to garner mass attention was posted on the 5th and accumulated almost 60k views. That’s almost 90% of the total view count from posts with PierrePutin hashtags that day. It claims Poilievre and the Conservative party are acting in the interest of the Russian government by showing a picture of the Russian flag with the Canadian Conservative Party logo in the middle.

Image 3: The first post containing a hashtag in the PierrePuting category (here #PierrePutin) to garner a significant amount of views.

A day later, on the 6th, a post was shared with the following text:

“Pierre PoiLIEvre's refusal to go through NSICOP security clearance screening is looking more and more plausibly connected to the Russian influencer disinformation and election interference campaign in the US and Canada #RussianInterference #PierrePutin #cdnpoli #elxn45”

This post, accompanied with an image of Poilievre in Parliament, collected around 68K views, 20% of the total views that day. While many posts made vague claims about Poilievre’s ties to the Russian government, others, like this one, were more explicit and precise in their accusations.

Finally, the top 6 most viewed PierrePutin posts occurred between Sept. 7th and 8th, with the most viewed post occurring on Sept. 8th and garnering almost 150K views (see Image 4). The post quotes a post sharing a video of a well-known political journalist criticizing Poilievre. The accusations here are more specific than the prior ones (e.g. Image 3) but they remain quite vague. This post appears to be generated as a way to join the conversation criticizing Poilievre more than hypothesizing explicit links with the Russian government and the Tenet Media right-wing influencers.

Image 4: The post with a hashtag in the PierrePutin category (here #PierrePutin) which got the most views in our period.

While #PierrePutin and other hashtags categorized as PierrePutin are not new, they garnered a lot of attention and use in the days following the Tenet Media story release. Many of the highly viewed posts contributed to the spread of the hashtag category made vague accusations or hypothesized about potential links between Poilievre, the Canadian Conservative Party, and the Russian government.

Was the emergence of the #PierrePutin manufactured?

To answer this question, we sought out to identify if the emergence was organic (i.e. follows a natural evolution of engagement and attention) or manipulated (i.e produced by bots or some other inorganic behavior).

First, we looked at the relationship between the number of posts with PierrePutin hashtags posted every six hours in our range, the number of unique users who posted them, and the total view counts of those posts. Figure 6 plots the normalized cumulative distribution of the number of posts, the number of unique users, as well as the view count. On the same Figure, we show equivalent metrics for the group of pro-Trudeau hashtags which had received attention around the same time.

Figure 6: Normalized cumulative distribution of the total number of tweets, number of unique users, and sum of all tweet views over a 6 hour timespan of tweets containing pro-Trudeau hashtags (dashed lines) and PierrePutin hashtags (continuous lines).

Looking at Figure 6, we find that the number of posts and number of unique users who posted them have almost the exact same growth pattern throughout the entire period. We find little to no divergence between the cumulative distribution curves for the post count and number of unique users, i.e. we don’t see a gap as speculated above. This indicates that there was a more or less constant ratio of post count to unique users throughout the entire time span. Furthermore, comparing to the evolution of another group of pro-Trudeau hashtags which were trending in relation to the NDP announcement of the end of the “supply and confidence” agreement, we find that the relation between the daily view count, number of posts and number of unique users is very similar for both hashtag categories. We find that PierrePutin hashtags followed a lifecycle similar to that of other hashtags.

For the second portion of this analysis, we manually reviewed the accounts that garnered the most attention for each day over the time period for which we collected posts. While some of the accounts were very frequent posters, we do not suspect any of the accounts of being bots. The users posted a variety of different political opinions, used human-like language (e.g. slang and culturally aware lingo), and exhibited reposting and quoting behavior similar to that of other users. We find no evidence that the emergence of the hashtag was inorganic and conclude it was Canadian users that responded to the Tenet Media incident by attempting to paint Pierre Poilievre and the Conservative Party as being beholden to Russian interests.

Conclusions

The DOJ indictment was a huge step forward in exposing a Russian-led disinformation campaign. However, that some Canadians took the opportunity to argue that Poilievre is acting in Russian interest falls in line with Russia’s general goal of destabilizing democratic institutions abroad and is itself damaging. There is no evidence to date showing that Pierre Poilievre and the Canadian Conservative Party have any formal ties to the RT executives who funneled money, Tenet Media, or the influencers who received Russian money. While some members share similar opinions, this does not constitute evidence that they are under the influence of Russian propaganda. Unfortunately, the connection we have captured here, between Poilievre and the Canadian Conservative Party with the Tenet Media case, may have triggered some undesirable outcomes; specifically, it may have contributed to increased distrust in Canadian political parties, and polarization and division in Canadian society. While the phenomena we have captured appears to have no connection with the Russian state, ironically, it does appear that the ripple effects align with the Russian agenda, even if unintended.

Methodology

Posts about Tenet Media and the end of the Supply and Confidence agreement

To identify posts about the Tenet Media story we looked for (case insensitive): “tenet,” “doj” and “indictment” and (“influencer” or pundits or “russia”), “influencers” and (“russia” or “russian”), “foreign” and “interference,” along with the names and Xtwitter handles of those involved (dave rubin, tim pool, lauren southern, benny johnson, matt christiansen, tayler hansen, liam donavan, kalashnikov and afanasyeva). For the supply and confidence stories, we looked for “supply and confidence,” “press conference” and “jag,” “press conference” and “singh,” “press conference” and “ndp,” “agreement” and, again, one of either “singh,” “jag” or “ndp,” “non-confidence” and “agreement,” “early election” along with “government” and “fall.” For keywords for both stories we also included their french equivalents (e.g. (“dja” or “département de la justice”) and “acte d’accusation” and (“influenceur” or “russie” or “russe”)).

List of hashtags collected from posts mentioning “Tenet Media”:

'pierreputin', 'neverpoilievre', 'pierrepoilievreisarussianasset', 'nevervoteconservative', 'pierrepoilievreislyingtoyou', 'istandwithtrudeau2025', 'pierreputinevre', 'trudeau2025', 'cdnmediafailed', 'nevervoteforpoilievre', 'freedumbconvoy', 'womenagainstpoilievre', 'poilieverislyingtoyou', 'putinpierre', 'nobodylikespierrepoilievre', 'istandwithtrudeau', 'pierreislyingtoyou'

Hashtag categorization for the 100 most used politically charged hashtags:

PierrePutin = "pierreputin", "pierreputinevre", "pierrepoilievreisarussianasset", "pierreputinresign", "pierrepoutin", "pierrepoilievreisaforeignasset", "putinpierre", "pierrepoilievreisaforeignasset", "pierrepoutin", "comradepierrepoilievre", "neverpierreputin", "peirreputin"

pro-Trudeau = "istandwithtrudeau", "istandwithtrudeau2025", "trudeau2025", "canadiansstandwithtrudeau", "istandwithttrudeau", "therearemillionsofus", "trustinjustin"
anti-Poilievre = "neverpoilievre", "nevervoteforpoilievre", "womenagainstpoilievre", "pierrepoilievreislyingtoyou", "pierrepoilievreislying", "pierrepoilievreisweird", "pierreislyingtoyou", "pierrepoilievreresign", "pierrepoilievreisunelectable", "pierrepoilievreisafascist", "pierrepoilievreisamisogynist", "sickofpoilievre", "pierrepoilievreisunelectable", "peepee", "conmansenseparty", "nobodylikespierrepoilievre", "pierreisnotmycanada", "pierrepoilievreisafraud", "pierrepoilievreisaliar", "pierrepoilievreismaga", "pierrepoilievreispathetic", "poilievreisunfitforoffice"
anti-Conservatives = "nevervoteconservative", "nevertrustaconservative", "nevertrustaconservatives", "neverconservative", "nevercpc", "republiconpartyofcanada", "maplemaga", "neverevertrustconservatives", "neverconservatives", "nevertrustconservatives"
anti-Singh = "selloutsingh"
anti-Freedom Convoy = "freedumbconvoy", "freedumbconvoy"
anti-Doug Ford = "dougfordiscorrupt", "dofomustgo", "fordfailedontario"
anti-Danielle Smith = "daniellesmithisunfittolead", "callanelectiondaniellesmith", 'daniellesmithisaliar', "daniellesmithiscorrupt", "danithedictator"

Anti-Cdn Media = "cdnmediacolludes", "cdnmediacorruption", "cdnmediafailed", "cdnmediafails"

Footnotes

(1) We calculate total engagement by summing tweets’ likes, retweets, quotes, replies and views.

Incident Update 7︱Don’t “Russian” to conclusion: An X analysis of downstream conversations in Canada