Veganuary and the vegan sausage (t)rolls: conflict and commercial engagement in online climate-diet discourse

The paper investigates how corporate participation in social media activism intersects with and affects vegan advocacy, using Veganuary 2019 on Twitter and the Greggs Vegan Sausage Roll (GVSR) incident as focal points. It addresses two aims: (i) to examine the character and effectiveness of corporate involvement in veganism, and (ii) to describe and explain the content and sentiment of the vegan culture wars. The study positions Veganuary within broader debates about the mainstreaming of veganism, corporate co-option, and the polarised backlash—often aligned with left-right political identities—around climate-diet discourse online.

The authors review scholarship on the mainstreaming of veganism as an environmental, health, and animal-welfare practice and its alignment with younger, urban, wealthier demographics. They discuss the rise of corporate or Big Veganism and critiques from traditional vegan activists and food justice scholars who argue that corporate-led plant-based products risk diluting radical vegan praxis and reproducing neoliberal consumerist solutions that leave power structures intact. Counter-arguments from meat and dairy interests and right-leaning voices are summarised, including claims about the cultural and ecological role of livestock (regenerative agriculture, sustainable intensification). Prior studies of online climate-diet debates (EAT-Lancet, IPCC) show polarisation, toxicity, and the emergence of pro-meat hashtags (e.g., #yes2meat). The literature indicates that discourse often spikes around scientific reports; this study extends that work to a commercial, consumer-facing context (Veganuary and GVSR). The authors clarify their working definition of veganism aligned with Veganuary’s practice-based emphasis (animals, environment, health) and note the complexity and diversity within vegan activism.

Data collection: Using Twitter’s Search API as available in 2019, the authors queried vegan, veganuary, and veganuary2019 four times during January 2019, merged results, and removed duplicates, yielding over 460,000 tweets. Tweets were primarily from the UK with additional posts from multiple countries. Ethics: Approved by the University of Oxford’s Social Sciences and Humanities Interdivisional Research Ethics Committee. The study adheres to norms of anonymising non-public individuals and only directly naming public figures, corporations, and organisations. Retweet network construction: The authors built an undirected, weighted retweet network connecting user A to user B if A retweeted B (weights equal total retweet counts between pairs, summed bidirectionally). First-level leaves (users with a single connection) were removed to highlight more active interactions. Community detection was performed with the Louvain algorithm; modularity and assortativity were computed to characterise polarisation and homophily. Rationale is provided for privileging retweets over likes, replies, and mentions, given retweets’ stronger signal of information spread. Community labelling: For each community with >1% of total retweets, a random sample of 200 tweets was manually coded. Initial coding used pro-Veganuary, anti-Veganuary, and commercial engagement, then refined into richer narrative labels (e.g., Veganuary-specific promotion vs generic vegan promotion, GVSR negative/positive, conspiracies). Communities were named by the dominant label in each sample. Audience overlap analysis (projection): To assess whether commercial audiences overlapped with activist audiences, the authors constructed a bipartite network linking top producer accounts (most retweeted) and their consumers (retweeters). They projected onto the producer layer, connecting producers with shared audiences, and applied community detection. The resulting monopartite network (141 nodes, 996 edges) was analysed for modularity (0.54) and composition by the original retweet-network communities to infer audience separation or overlap. Robustness considerations and methodological details reference established bipartite projection methods.

- Discourse structure: Twelve large retweet communities clustered into four themes: Veganuary activism (Core Support), antagonism (e.g., Access/Trolls, Piers Morgan, Conspiracies), commercial engagement (Veganuary-specific and generic promotions), and GVSR-focused discussion (GVSR- and GVSR+). Examples: Core Support comprised 14% of users and 22% of retweets; GVSR- 15% of users and 14% of retweets; GVSR+ 14% of users and 12% of retweets; Access/Trolls 10% of users and 18% of retweets (Table 1). - Content dynamics: Activists promoted benefits of veganism, shared resources, and defended the movement, with some intra-vegan boundary policing. Antagonist communities included critiques of elitism and accessibility, insults toward vegans, conspiratorial claims about corrupt climate-diet science, and politicised attacks linking veganism to left-wing agendas. - GVSR flashpoint: The GVSR controversy generated two major communities. GVSR- contained aggressive and sometimes violent rhetoric from both critics and defenders; GVSR+ was more lighthearted, supporting trying the product and mocking outrage. The incident amplified culture-war dynamics and drew in right-wing influencers and groups. - Audience projection: The producer projection network (141 nodes, 996 edges) exhibited strong audience separation (modularity 0.54). Commercial communities’ audiences largely did not overlap with activist audiences, indicating that commercial engagement did not drive their followers toward Veganuary activism content. Exceptions were tied to the GVSR, whose audience overlapped with other parts of the discourse due to controversy rather than activism. - Antagonist audience patterns: The negative GVSR community (GVSR-) remained isolated from other antagonist audiences. Conspiracies clustered with much of the Piers Morgan community; a distinct subset aligned with US-based QAnon supporters formed a separate audience cluster. - Impact note: Despite backlash, Greggs reported a 58% sales increase in H1 2019, highlighting commercial benefits even as activist engagement gains were limited.

The findings address the research questions by showing that corporate participation in Veganuary 2019 generated substantial content but did not connect commercial audiences to activist discourse. Audience projection indicates that commercial campaigns generally served insular product-promotion goals rather than bridging to activist messaging. The GVSR was a notable exception in visibility, but its overlap stemmed from controversy and politicisation, not increased vegan advocacy. The discourse displayed pronounced culture-war characteristics: anti-vegan narratives tied to right-wing politics, climate denialism, and conspiracy theories, with risks of drawing in radical groups (e.g., QAnon) and normalising antagonistic rhetoric. These dynamics can distract from activist goals and potentially escalate polarisation. The study suggests that for corporate engagement to meaningfully support activism, closer coordination and accountability mechanisms are needed to align promotions with movement values and to channel audiences toward activist resources.

Commercial engagement with Veganuary 2019 on Twitter largely failed to connect corporate audiences with vegan activists or deepen public understanding of veganism’s ethical and environmental motivations. While companies participated and benefitted commercially (e.g., GVSR visibility), their efforts typically remained siloed and, at times, undermined perceived authenticity of the movement. The GVSR controversy exemplified how culture-war triggers can dominate discourse, drawing attention away from advocacy and attracting antagonistic and radicalised participation. Future research should examine multi-year and multi-platform dynamics, identify strategies to counteract polarisation, investigate the roots of vegan antagonism (psychological, cultural, communicative), and develop practical frameworks for corporate-activist coordination that builds trust and channels engagement to movement objectives.

- Temporal scope: Analysis focuses on January 2019; results may not generalise across years with differing political or media contexts. - Platform scope: Twitter-only data may not represent broader climate-diet discourse across other social platforms or offline contexts. - Interaction signals: Method prioritises retweets; likes, replies, and passive exposure were not incorporated, potentially missing other influence channels. - Network construction: First-level leaves removed, which may exclude low-engagement but potentially meaningful interactions. - Causality: Audience projection detects shared audiences but cannot infer the sequence or directionality of engagements. - Generalisability: Predominantly English-language and UK-centric sample limits global representativeness.