Affordances of Wartime Collective Action on Facebook

Abstract

This paper relies on the conceptual framework of affordances to study collective action, focusing on the under-researched area of wartime collective efforts. Using the case study of collective action during the war in Ukraine, the paper analyses a sample of the most successful Facebook profiles engaging in collective action to support the military and humanitarian efforts in Ukraine and uncovers the various approaches to the organisation of collective action. The paper explores the common features and differences between activists, including approaches to presentation, user engagement, interaction, and reporting. Two groups of profiles are identified based on how they actualise Facebook affordances: organisational and individual profiles. The two groups were found to clearly differ in their approach to user engagement, posting patterns and the use of specific Facebook functions. The findings have implications regarding the organisation of collective initiatives by different actors and contribute to a better understanding of the unique challenges of wartime CA.

1. Introduction

The phenomenon of collective action (CA) in the digital realm has captivated the attention of academic researchers since the 1990s, who pointed out the reduced costs of organisation as a significant advantage of the new medium (Bonchek, 1995). Over the years, the advent and widespread adoption of social media platforms have led to a surge in scholarly interest, sparking numerous investigations into the relationship between online spaces and collective mobilisation (Castells, 2012; Tufekci, 2017). Initially, some scholars remained sceptical about the role of social media in facilitating CA, dismissing their significance as “clicktivism” or “slacktivism” (Morozov, 2009; Skoric, 2012). With time, however, the mounting evidence of the role of social media in the Arab Spring protests (Lotan et al., 2011), the Ukrainian protests of 2013–2014 (Ronzhyn, 2014), the election of Donald Trump, and Brexit’s Leave campaign (Hall et al., 2018) have compelled researchers to reevaluate their stance on the topic.

The year 2022 provided an unfortunate and unique setting for studying online CA. With the start of the full-scale Russian invasion of Ukraine on 24 February 2022, there has been an incredible rise in volunteering in Ukraine. Thousands of people joined existing volunteer organisations or created new ones to supply the Ukrainian army and support the numerous victims of war (Ioanes, 2022). Alongside the rise in offline volunteer movements, there has also been a corresponding mobilisation online, with virtual communities coming together to gather support for Ukraine’s forces and citizens.

To date, there has been no social media analysis focused on CA in Ukraine. Moreover, most of the research on social media during wartime is focused on media framing (Badran & Smets, 2018; Makhortykh & Sydorova, 2017) and the use of social media as a weapon (Nissen, 2014; Prier, 2020). Wartime social media activism also remains relatively unexplored: previous papers include explorations of the role of video activism during the Syrian conflict (Andén-Papadopoulos, 2020), protest activism during military occupation (Wulf et al., 2013), and military recruitment on social media (Krona, 2019). All these papers, however, do not explore CA beyond the organisation of protests, leaving a major gap in the research, which this paper aims to address.

In our research, we focus on the case of the war in Ukraine and the use of social media for the organisation of CA during the first year of the Russian invasion. Social media played a significant role in the protests that led to the overthrow of the Yanukovych regime in 2014 (MacDuffee Metzger & Tucker, 2017; Ronzhyn, 2014), and many of these social media profiles have continued to operate during the subsequent Russia–Ukraine conflict in the east of Ukraine (Worschech, 2017). The attack on February 24, however, marked a sudden and unprecedented increase in activity.

We use the conceptual framework of affordances as a way to frame and explore the different uses of the platform for supporting the war effort and activism. Affordances provide a useful theoretical lens for studying social media, as they acknowledge the complex interplay between the users and the system rather than focusing solely on the system’s functions (Nagy & Neff, 2015).

The remainder of the paper is structured as follows: Section 2 and Section 3 discuss two theories employed in the paper, CA and affordances, and describe their relation. Section 4 presents the methodological approach. Section 5 and Section 6 discuss the results of the analysis and the implications of the findings. The paper ends with the conclusions section, which summarises the research findings.

2. Collective Action

Collective action refers to the activities of a collective in pursuit of a common goal (Wright, 2009). It is often discussed in contrast to individual action, aimed at reaching personal goals. Mancur Olson’s (1965) work on CA laid the foundation for the study of CA in economics and political science. Subsequently, scholars have built upon his ideas and expanded the study of CA to various fields, including social movements, political participation, and public policy (Klandermans, 2015; Pecorino, 2015).

More recent research has focused on the role of technology in CA. The proliferation of the Internet has undoubtedly changed some of the conditions in which CA is organised: specifically, it has lowered communication and organisation costs (Bimber et al., 2005), mitigating or eliminating the disadvantages of larger groups (Castells, 2012; Lupia & Sin, 2003). Social media, in particular, has played a significant role in most of the protest action and bottom-up activism over the recent decade (Lee, 2015; Ruijgrok, 2017).

Underlining the great transformative power of the networks, Bennett and Segerberg (2012) suggest that the Internet and online interaction established a different logic altogether: the connective action, which in contrast to CA, is not associated with high levels of organisational resources and the formation of collective identities. In contrast, connective action is characterised by a lack of organisational coordination and personal expression shared over social networks (Bennett & Segerberg, 2012). On social media, small contributions through sharing, liking, or following can sometimes scale up and significantly contribute to the CA cause (Margetts et al., 2014).

Both warfare (Lopez, 2020) and military resistance have been discussed as a CA problem (Kalyvas & Kocher, 2007; Wood, 2003). Nonetheless, the war in Ukraine has several unique properties absent from previous large-scale conflicts and wars: the large numbers of people affected by the war directly, the high penetration of the Internet and social media within the affected region, and the willingness of individuals and organisations in Ukraine and abroad to contribute financially to the war effort. In this context, social media function as efficient infrastructures for organising and sustaining donation-based collective action, enabling individuals to support Ukraine through monetary contributions, fundraising campaigns, and direct financial transfers to trusted actors. However, collective initiatives in wartime also confront serious constraints: the volatility of conflict zones, pervasive disinformation campaigns, censorship or content moderation, and the fragility of trust under high uncertainty (Li, 2025; Pierri et al., 2023). Platform governance and moderation, especially under emergency or crisis conditions, may impede visibility or penalise ambiguous content, complicating fundraising appeals. Moreover, sustaining engagement becomes harder as audiences fatigue or shifting priorities divert attention (Marshall et al., 2024). While recent work on wartime digital mobilisation (e.g., Ye et al., 2023) has begun exploring these challenges, our study contributes by systematically linking affordance use to behaviour under these constrained conditions.

3. Social Media and the Theory of Affordances

When discussing CA in social media, it is necessary to employ a coherent framework for understanding the use of different functions and technologies of the platforms. The theory of affordances can be such a useful tool for conceptualising specific aspects of social media that either contribute to or constrain CA. The concept of affordance was initially coined by Gibson (1977, p. 75) who defined it as a “unique combination of qualities that specifies what the object affords us”. The concept was applied to different fields, most notably product design (Norman, 1988), and subsequently to communication (Hutchby, 2001), as a way to describe how technologies enable and constrain specific actions of their users. Applied to social media, affordances can be defined as “perceived actual or imagined properties of the technological products, emerging through the relation of technological, social and contextual, that enable and constrain specific uses of the platforms” (Ronzhyn et al., 2022, p. 14). Such properties can be general: persistence, anonymity, or visibility (Evans et al., 2017; Treem & Leonardi, 2013) or specific: content association, feedback directness, and audience transparency affordances related to self-presentation on social media (Devito et al., 2017). Any social media platform has a certain number of affordances, and while all the affordances of a platform are available to each user, the actualisation of affordances depends on two factors. Firstly, the perception of a user: affordances are perceived properties, so to actualize a specific affordance a user needs to be aware of it. Secondly, the agency of a user: ultimately the decision to actualize or not to actualize an affordance is left to each user, and it is made based on context (objectives of the user, their audience, relations to other users). Explicitly considering these interactional aspects of technology use, the theory of affordances allows us to avoid technological determinism (Hutchby, 2001).

Affordances theory has been prominently used in management and organisation studies research, while its application in CA research is limited to case studies of specific movements. Such studies also disproportionately focus on the identification of social media affordances. For example, Sæbø et al. (2020) and Harindranath et al. (2015) each identified nine affordances supporting CA; Vaast et al. (2017) distinguished between individualised, collective, shared, and connective affordances of social media; and Zheng and Yu (2016) outlined two general SM affordances for CA: agenda setting and framing.

While these studies showed the usefulness of affordances as an analytical tool in this domain, they did suffer from some drawbacks. Firstly, they tend to conflate affordances with the actualisation of affordances, that is, mixing what the media afford with behaviours produced through the actualisation of the affordances (see, for example, Evans et al. (2017) for discussion on that). Secondly, the abovementioned papers treat social media as a monolith phenomenon, failing to recognise the differences between SM platforms, a common issue in the early years of digital research, which should be avoided now with the platformisation of the Internet (Poell et al., 2019).

With this paper, we aim to address two research gaps. The first one, presented in the previous section, is related to the understanding of social media-based collective action during wartime. The second one, justified above, is on the use of affordances as an analytical tool describing the use of social media for realising CA initiatives. To achieve these two goals and to avoid the pitfalls plaguing the previous studies, in this research we focus on one social media platform (Facebook) and clearly distinguish between affordances and behaviours that result from the actualisation of affordances. The suggested model of relations between the core concepts on Facebook is presented in Figure 1. It considers affordances offered by the platform, behaviours as the results of actualization of affordances, messages conveyed to the audience through platform behaviours, and finally, the response of the audience in form of user engagement with content created by CA profiles. Audience engagement with the content may inform the future behaviour of the CA profile owner, who may adjust their social media strategy.

As a starting point, we take two sets of social media affordances: four general Facebook affordances identified by Treem and Leonardi (2013) and scrutinised by Ronzhyn et al. (2022), and three affordances formulated by Etter and Albu (2021) specifically for the case of CA in social media. By analysing the affordances and their potential to be measured we came up with a final list of six affordances used in this paper. We used the actualization of these affordances by profiles (public representations of social media users, who can be organisations or individuals) to illustrate the differences in the platform’s use. The six affordances and the suggested ways to measure them, as well as the expected variance between profiles in the sample, are shown in

Table 1. Facebook affordances and ways to measure the degree of their actualisation.

Facebook Affordance	How to Measure	Expected Variance Between Profiles
General Facebook Affordances (Treem & Leonardi, 2013)
Visibility	Number of posts/post types	High
	Post engagement
Persistence	Number of deleted posts	Low
	Age of profile
Association/Interlinking	Use of tagging and post sharing functions	High
	Use of hashtags
Editability	Frequency of editing published posts	Low
CA-Specific Facebook Affordances (Etter & Albu, 2021)
Assembling	Cannot be measured from public data	High
Augmenting	Post engagement	High
	Use of hashtags
	Cross-platform engagement

.

Visibility refers not only to the accessibility of the published content but also to how visible the content is relative to the content produced by other profiles. There is likely a high degree of variation between profiles based on their number of followers and their behaviour. Here, it is worth considering the role that non-human agents play within the social media ecosystem by influencing its functioning and what the media afford their users (Sharma et al., 2016). Most notably, social media algorithms mediate the visibility affordance by affecting what is visible to or hidden from a particular user. In this way, algorithms may work against the intentions of activists by hiding calls for action and other content meant to reach as wide an audience as possible (Bucher, 2012; Bucher & Helmond, 2017).

Persistence refers to how the created content remains accessible and immutable over time. Persistence allows the content to be searched, replayed, and restructured (Erickson & Kellogg, 2000). For CA it allows an audience to observe the continuity of action of the organisations, serving as evidence of good conduct for new followers, and thus helping to build the organisation’s reputation. Variance in the actualisation of persistence is likely low, as the removal of previously published posts is rare. Editability is the flipside of persistence; actualising it allows users to modify published content. Facebook keeps the history of edits for posts, allowing the persistence of all versions of the content (at least all the textual versions, as the deleted images are not stored). Editability plays a role in Facebook’s moderation practices, as, in some cases, users are allowed to keep a publication that infringes upon Facebook rules, given it is edited to conform.

Association is an affordance through which Facebook gives its users the capability of demonstrating connections to other users. This capability, realised through sharing posts and the function of tagging, allows a user to include a list of other users (pages or individuals) at the top of their post. The use of the tagging function is not universal and benefits less-popular profiles more than established ones, thus we predict that the variance here is high.

The Interlinking affordance outlined by Etter and Albu (2021) is encompassed by association as the corresponding behaviours and therefore indicators are the same for both affordances.

Assembling is an affordance for collecting and collaboratively working on information, including the co-creation of press releases or reports.

Finally, Augmenting refers to the possibility of attracting more attention to the campaign by amplifying outreach and exposure. It can be actualised through post reactions and shares, as well as the use of hashtags and use of other platforms as a way to garner more support.

Among the specific affordances, Interlinking (within a more general Association) and Augmenting can be discussed with the collected data, while assembling refers more to the internal operations of the profiles and can be only inferred indirectly. Moreover, the assembling affordance of Etter and Albu (2021) principally describes the cooperation of individual activists, while in our case there is primarily many-to-one cooperation, with a fundraising profile as a central node in CA. Thus, the assembling affordance will not be discussed further in the paper; however, it might be the focus of a future paper employing interview or survey methodology.

Engagement and interaction metrics on social media can be understood as behavioural manifestations of intentional affordance actualization. When users intentionally activate affordances such as visibility, association, and augmenting, these actions manifest as measurable engagement indicators (likes, shares, comments). Importantly, engagement also represents the fundamental link between individual behaviour and collective action: it is through the cumulative effect of individual acts of engagement that collective action emerges and gains momentum. In this sense, engagement serves not only as a proxy for the intentional actualization of affordances but also as the observable expression of participation in collective processes: there is no collective action without individual engagement.

4. Individual and Organisational Profiles

When affordances are actualised, they produce behaviour, which can be measured through various social media metrics that concern the use of different platform functions. Again, while affordances are available to all users, they can be actualised to various degrees and with different frequencies. There might be various reasons why certain profiles favour one affordance, while others prefer the other. One obvious starting point is the relation with the intended audience. This relation, which can be summarised as trust or reputation, is an important factor for users, who are deciding to donate money or time to a non-profit organisation or a fundraising campaign on social media (Wymer & Čačija, 2022).

There is some evidence of differences in how trust and reputation are built and maintained by individual versus organisational profiles. Previous research suggests that organisations focus more on organisational performance and social accountability (Dowling, 2004) to build reputation. Transparency in their operations and accountability are important factors of legitimacy for fundraising organisations (Zhou & Ye, 2021), positively influencing trust (Farwell et al., 2019) and donation behaviour (Becker, 2018). For individual profiles, a different aspect is more important: defined as personality or authenticity (Dowling, 2004). Driessens (2013) suggests the term “celebrity capital” to denote the value acquired in this way, aiming at fostering trust (Sulaeman, 2017). This difference between organisations and celebrities likely leads to different strategies on Facebook.

To examine these assumptions, we formulated the following hypothesis: the use of Facebook by CA initiatives depends on the owner of a profile: organisational and individual profiles behave differently (in terms of engagement strategies, posting patterns and use of platform functions) and thus focus on different technological affordances [H1].

The goal of collective action initiatives in this context is to mobilise as many resources as possible from their followers. Because direct measures of success (e.g., financial contributions, supplies, number of volunteers) are unavailable, engagement rate serves as a behavioural proxy for the effectiveness of these efforts. It reflects followers’ participation and responsiveness to the profile’s mobilisation attempts. Thus, we state [H2] that differences in affordance actualisation are systematically associated with variations in engagement outcomes across profiles [H2].

Finally, as revealed in preliminary interviews and previous scholarship on platform governance, the actualization of affordances is not entirely under user control but constrained by the moderation and algorithmic policies of social media platforms. Visibility, persistence, and augmenting affordances, in particular, may be limited or distorted by content restrictions and automated enforcement. Therefore, we propose a third hypothesis: the actualization of affordances in wartime collective action is constrained by the moderation policies and algorithmic governance of social media platforms [H3].

To test the hypotheses, we analysed a sample of eight Facebook CA profiles (including both organisations and individuals) and the content produced by these profiles over the first nine months of the war, examining the actualisation of various Facebook affordances to determine if there are meaningful differences between the two groups.

Addressing the hypothesis will contribute to the more general objective of the paper: examining the wartime CA landscape on Facebook and describing the organisation of the initiatives, placing it within a broader discussion of online CA. Central to this investigation is an exploration of how these initiatives leverage the affordances provided by the Facebook platform to achieve their ultimate goal. Analysing and characterising these diverse approaches will provide valuable insights into the dynamic and evolving nature of online communication during times of conflict and contribute to the literature on technology-mediated mobilisation and the role of social media platforms in shaping CA.

5. Methods

In this study, we adopt a combination of qualitative and quantitative approaches to analyse the profiles organising CA on Facebook. We aimed to include all popular public profiles (pages and groups with at least 10,000 followers and public visibility) actively engaged in CA. The profiles were found in December 2022 through keyword search (using “donate”, “fundraising” + “(bank) account”, “PayPal” keywords in Ukrainian) and by following links and tags in the posts of already identified pages. As the number of such profiles is low and they are quite visible, we have likely identified all the profiles fitting the inclusion criteria. One popular profile (of a writer, Dorje Batuu) was included in the initial sample but has been deactivated by the owner since and is thus excluded from our sample. The full sampling process is shown in Figure 2.

We analysed 8 Facebook profiles, listed in

Table 2. Facebook profiles analysed (follower counts—November 2022).

Code	Facebook Profile	Type	Followers
O1	Come Back Alive	Organisation	3.5 M
C1	Serhiy Zhadan 1	Celebrity	162,362
O2	Razom for Ukraine	Organisation	48,803
OC	Fond Serhiya Prytuly	Organisation/Celebrity	36,422
O3	Armiya SOS [group]	Organisation	33,467
O4	Hospitallers Paramedics	Organisation	32,820
O5	Pidtrymai Armiyu Ukrayiny	Organisation	11,987
C2	Andriy Lyubka	Celebrity	11,442

¹ Transliteration is used for some of the pages with names in Cyrillic script.

(and henceforth mentioned by their codes for brevity), focused on fundraising for the army and/or humanitarian purposes in Ukraine and having at least 10 thousand followers. CrowdTangle1 was used to collect information from the profiles and extract posted content (3892 posts), together with additional quantitative information reported alongside the qualitative findings. The sample includes all the public posts for the listed profiles within the nine-month period from 23 February 2022 (one day before the war) to 23 November 2022. Profile statistics and trends were collected in November 2022, thus any posts deleted by that point are not included in the dataset. While the one deleted profile was active during the analysed period and had 27.5 thousand followers at the time of deletion, it is not included in our sample.

In addition to a significant variation in the number of followers, there are some other important differences between the profiles. The most significant difference related to the hypothesis is the ownership of a profile: some profiles are online representations of individuals (C1, C2), while others are organisational profiles (O1, O2, O3, O4, O5). One profile (OC) is an organisation; however, it has a strong emphasis on the leading person (for whom the Fund is named). The difference in ownership is not only formal but is also reflected in how communication is organised: both in terms of the content of the messages shared and in the way the profile interacts with its followers.

Fundraising is by far the most common action organised on social media in a time of war; all of the analysed profiles engaged in fundraising on Facebook. Additionally, profiles also collected and distributed supplies (often large donations from companies or organisations). Some profiles (O4) also organised education and training activities, as well as the online and offline activism (O2). Most profiles already existed before the war started. In our sample, only one profile (OC) was created after February 24. Charity organisations existed because there was ongoing warfare in the east of Ukraine, while celebrity profiles existed as a means for a famous person to get in touch with their audience. Still, all profiles grew their audience more rapidly after the start of the war. Most profiles achieved their most substantial growth in the first week of the war (23 February–2 March); the mean increase in followers was 2.71%, with the highest increase being 8.78%.

To classify the posts, we used a combination of automated and manual tagging: the posts were first tagged according to the keywords and subsequently reviewed by the researchers to ensure that the assigned tags were correct. We focused on several areas related to platform affordances discussed in Section 3, depicted in the conceptual model in Figure 1: posting patterns, post content, reporting. We assigned tags denoting posts that contained reports, calls to action, and personal details as well as various other topics (e.g., everyday operation of funds, personal photos, reporting photos, and others.

As a supplementary method to capture aspects of Facebook use that cannot be assessed through content analysis alone, we conducted short semi-structured interviews with the individuals responsible for managing the CA profiles. The aim was to identify the actualization of affordances that remain invisible in the content (particularly editability), as well as practices related to moderation and algorithmic interaction. The themes and guiding questions used in the interviews are provided in Appendix A.

Interviewees were recruited through direct messages sent via the official Facebook pages of all eight profiles included in the study. Eligibility was limited to those directly involved in profile management or communication strategy. Three profiles (37.5%) agreed to participate. In two cases, the respondents were page administrators, and in one case, the communication manager of the organisation. All interviews were conducted in Ukrainian (the interview guide was translated for inclusion in the Appendix A), lasted approximately 15–25 min, and were carried out via secure online calls or, in one case, via voice messages on an encrypted messenger. With participants’ explicit consent, the interviews were recorded for transcription purposes. Participants received detailed information about the project and provided written informed consent prior to the interview; no compensation was offered. Owing to the sensitive context, the recordings were deleted after transcription and coding.

The analysis followed a primarily deductive approach, guided by the affordance framework: responses were coded manually using a predefined set of affordance categories (editability, visibility, persistence, association, and augmenting). Coding was performed by the first author and validated by two other co-authors; discrepancies were discussed until consensus was reached.

The interview findings enriched our understanding of how CA profiles interact with Facebook’s algorithmic systems, moderation mechanisms, and potential adversarial effects such as shadow banning—insights that directly inform the examination of the third hypothesis (H3) concerning the constraining effects of platform governance on affordance actualization.

6. Results

6.1. Post Types: Calls for Action, Reports and Updates

Most posts published by the profiles are either calls to contribute or reports on the use of the donated funds. We used tags “call for action” and “report” to categorise the complete sample of posts of the profiles. Based on the ratio of reports to calls, two groups emerge (

Table 3. The ratio of reports to calls to action in the sample.

Profile	O1	C1	O2	OC	O4	O5	C2	Total
Calls	437	70	183	77	308	26	26	1128
Reports	378	155	130	196	129	21	94	1103
Ratio (R/C)	0.86	2.21	0.71	2.55	0.42	0.81	3.61	0.98

): one in which profiles publish more calls for action (O1, O2, O4, O5) and others that publish more reports (C1, OC, C2). Looking at the content of the posts, it is possible to observe further differences. Some profiles chose to include donation information almost in every publication (O1) while publishing very few posts for the exclusive purpose of encouraging donations. In contrast, during the first few months of the war, O2 was heavily involved in digital activism (orchestrating Twitter storms, petitioning) as well as in organising offline protests: thus, a significant bulk of posts were calls to attend the rallies or participate in dissemination of protest hashtags. O4 focused on fundraising and, while not providing many reports, published a lot of posts about the life of the organisation and their activity on the frontline as a medic battalion. While these posts were not reports per se, they did provide indirect evidence of the good use of collected funds and contributed to the authenticity of the profile. In this way, O4 is similar to both celebrity profiles (C1 and C2), which have a higher ratio of reports to calls, but do, however, publish a significant number of posts that relate to their day-to-day life.

The further difference between organisation and celebrity profiles is the level of detail provided in the reports. O1 provides the most detailed reports, also offering a search engine on their website where each donation can be viewed2. Other organisational profiles (O2, O4, and OC) provide less detailed but also very specific reports (i.e., with detailed sums collected). The length of reports also varied significantly (

Table 4. The average length of reports in the sample (in words).

Profile	O1	C1	O2	OC	O4	O5	C2
M length	157.55	70.3	153.8	111.1	110.4	182.8	6.9
Std. Dev	67.1	49.6	108.9	73.8	127	79.0	10.4

) with C2 as an extreme example providing visual reporting through photos with barely any words. All profiles relied heavily on visual content: 91.6% of all reports contained images, while a further 7.1% included video content.

In contrast to organisational profiles, C1 and C2 publish reports that are far less formal: their reports usually either show a picture of acquired equipment (often a car, as transport is the main focus of both celebrity profiles), or a report of collected funds, even with a total sum but little reference to when it was collected (in case of C1). Figure 3 shows examples of formal and informal reports.

Association/interlinking affordance is actualised through the use of hashtags and tagging. Tagging was successfully used by some profiles (O2, O3) as a promotion strategy: to increase the popularity of posts and attract new followers. Hashtags were used to connect a message to a wider discussion and to organise messages internally within one’s profile: acting as categories or tags for publications. Actualisation of the Association affordance contributes to the willingness to donate: interviews revealed that business donors are interested in being mentioned in the donation reports by the funds, particularly due to a very high number of followers of these funds, which dwarf the follower counts of most businesses in Ukraine.

Association/interlinking also allows users to show their performance by chaining the reports, either through hashtagging or using other methods: for example, C1 starts every report about a purchased car with a number (e.g., 40/100 in Figure 2), which both shows the total performance of the profile (40 cars bought already) and underlines an intention to continue (40 out of 100). Persistence of content allows a follower to search through the previous posts to see how the funds were used and thus boosts their confidence and trust in the profile.

6.2. Audience Engagement and Interaction

While the bulk of the content published by the profiles can be classified as calls to donate and reports of the use of donated funds, there are other types of content: shared posts from other profiles, news, and personal updates (celebrities’ creative life or the everyday activity from organisations). Although such publications do not contribute directly to the CA cause, they still can be useful indirectly: for keeping the audience’s attention, providing a human face for the cause, and promoting others working for the same objective. The percentage of personal posts is higher among the celebrity group (62% for C1, 23% for C2), while reports also often include personal stories and photos: underlining a more intimate connection with followers and giving them glimpses into the organisers’ personal lives (boosting the perception of authenticity).

To measure user engagement with profile publications, we use the interaction rate. The interaction rate indicates how many users have interacted with posts from the profiles (liked, commented, or shared) over a given month. The formula for calculating interaction rate is as follows: number of all interactions/number of posts/average profile likes × 100. The average interaction rate of the posts from the profiles in the sample for the whole period is between 0.24% and 5.24% (Figure 4), with marked variations both between and within profiles.

Some of these variations can be explained as a consequence of more devout followers starting to follow the profile early: they interact with the profile much more eagerly. Later, when people started liking the page as a sign of their solidarity or support, the interaction rate dipped in line with other profiles in the sample. A clearer picture is visible when we focus on the second part of the sample period (June–November 2022), shown in Figure 5.

The engagement rates are connected directly to the visibility and augmenting affordances. From the perspective of affordance theory, engagement and interaction rates can be understood as behavioural outcomes of affordance actualization. Thus, interaction rate serves as a proxy for the extent to which these affordances are successfully realised by the profile owners. Profiles with higher engagement are therefore not only more visible but also demonstrate greater effectiveness in converting the platform’s affordance potential into participatory user actions. Two groups of profiles can be identified based on interaction data: three profiles (C1, C2, OC) are above or near the interaction level of 2% (M = 2.98). C1 and C2 are profiles of celebrities, while OC is a charity organisation that also has a celebrity as its founder and speaker. O1, O2, and O3 are charity organisations that have much lower interaction rates (M = 0.46). O4 is a volunteer charity with an interaction rate between the two groups (with a spike going as far as 5% in April and May). Such high engagement rates can be explained by looking into the content published by the profile: O4 participated in the war directly as a medical battalion by the same name. Thus, they posted images and stories from the frontline, gathering more interactions, especially in the early months.

6.3. Post Content: Moderation and Algorithms

Facebook does not grant equal visibility to all posts. It employs an algorithm that favours certain posts while limiting or hiding the others (Bucher, 2012). While algorithms are somewhat opaque, they set rules that profiles have to follow to achieve maximum visibility of their content. Interviews revealed how following these rules constrains both the form and the content of publications, affecting both the CA message and how it is conveyed.

Content removal on Facebook is associated with two issues: hate speech and the publication of violent or sensitive images. Although in the first months of the war, the Facebook rules on hate speech were somewhat relaxed (Vengattil & Culliford, 2022), later they were enforced again, as noted by the CA profiles in our interviews.

For the profiles in our sample, there are inevitable trade-offs during the preparation and publication of the CA-related content. On one hand, the war-related content, including the depiction of weapons and certainly the use of weapons against enemies can be penalised, on the other, the whole point of activity of these profiles is aiding the war effort in Ukraine. As one of the interviewees noted, it is difficult to report on the acquired and supplied weapons, when you cannot show them in photos.

And while the weapon- and war-related content is not banned by the platform outright, it is penalised through shadow banning, reported by all the interviewees. Thus, all the interviewed profiles admitted to optimising and self-censoring the content to suit the demands of the Facebook algorithm, both in terms of content (avoiding formulation that might be considered hate speech and blurring sensitive photos) and in terms of form (e.g., editing and publishing videos through Facebook’s internal editor rather than YouTube). The biggest fund, O1, admitted to changing their posting behaviour over time and moving all the graphic videos or images to other platforms, keeping more neutral stories and photos. Such self-censorship is more prominent for organisational profiles compared to individual profiles.

Regarding the editability affordance, profile owners acknowledged that they had to remove photos sometimes (“several times in the last half a year”), mostly due to requests from people depicted. Despite their prior consent, due to the sensitive nature of the photos (e.g., fear that a place may be identified for a Russian airstrike), sometimes people ask them to remove the photos. In such cases, the profiles always comply. These findings directly support Hypothesis 3, illustrating how moderation policies and algorithmic governance constrain the actualization of affordances on Facebook.

7. Discussion

This study tested three hypotheses regarding how collective action profiles on Facebook actualize platform affordances under wartime conditions. The first hypothesis proposed that the use of Facebook by CA initiatives depends on the ownership of the profile, meaning that organisational and individual profiles behave differently in their engagement strategies, posting patterns, and use of platform functions. The results are consistent with this assumption. The analysis shows clear differences between these two groups: organisational pages rely on performance transparency and formal reporting to sustain legitimacy and trust, while individual profiles emphasise authenticity and emotional connection. These behavioural patterns correspond to distinct affordance strategies—organisations primarily actualize the persistence and association affordances to demonstrate accountability and maintain credibility, whereas individuals draw on visibility and augmenting affordances to express personality, reinforce authenticity, and foster audience identification. Given the small number of profiles analysed, these findings are interpreted as descriptive patterns rather than statistical generalisations. Nonetheless, they illustrate consistent and theoretically meaningful distinctions in affordance use and engagement strategies.

The second hypothesis proposed that systematic differences in engagement rates correspond to how profiles actualize the available affordances. This assumption is also supported. Profiles that made fuller use of the visibility and augmenting affordances—particularly celebrity profiles and the hybrid case—achieved substantially higher interaction rates (M ≈ 2.98%) compared to organisational pages (M ≈ 0.46%). The volunteer medical battalion (O4), which combined visibility and augmenting through direct frontline storytelling, achieved engagement levels that were, for a time, comparable to those of celebrity profiles. These findings suggest that engagement functions as a behavioural indicator of affordance actualization, linking the underlying strategies of transparency and authenticity to measurable interaction outcomes rather than treating engagement as an external metric of popularity.

The third hypothesis addressed how platform-level moderation and algorithmic governance constrain affordance actualization. Evidence from the interviews supports this assumption: all interviewed profiles reported instances of shadow banning, content removal, and algorithm-driven visibility penalties. These constraints compelled actors, particularly organisational profiles, to self-censor and strategically modify content to avoid sanctions. Hence, moderation policies shape the extent and form of affordance realisation, underscoring that affordances are not purely technical possibilities but contingent on platform governance.

Taken together, these patterns illustrate two coexisting affordance strategies for mobilising resources on social media: transparency-through-persistence (organisations) and authenticity-through-augmenting (individuals).

While organisations pay more attention to organisational performance (more frequent reports) and accountability (more detailed and specific reports), individual celebrities focus heavily on the personality aspect. Authenticity is seen as an indicator of reputation (Fombrun & Van Riel, 2004) and arguably it is more important for celebrities than organisations. Celebrity profiles have a greater number of posts unrelated to the CA efforts but describing the life and activity of the profile owner (which often still heavily features the war), while also self-censoring less to conform to the platform guidance. Such posts contribute to the CA indirectly by keeping the audience engaged and interested (as evidenced by the engagement metrics). The differences between the two main groups of profiles, discussed in the previous section, are summarised in

Table 5. Differences between the two groups of profiles.

Facebook Affordance	Actualisation of Affordance
	Organisational Profiles	Celebrity Profiles
Visibility	Very high number of reports to calls to action.	High number of reports to calls to action.
	Low post engagement per user	High post engagement per user
	Post content optimisation	Post content optimisation
Persistence	Use of content persistence to boost transparency and show performance	Use of persistence to boost authenticity and show transparency
Association/Interlinking	Use of post sharing functions to attract and promote donors	Use of tagging
	Use of hashtags to link the reports	Greater personal engagement with users
Editability	Low number of edited posts	Low number of edited posts
Augmenting	More frequent use of additional functions (e.g., live videos)	Use of emotion-evoking imagery
	Use of cross-platform engagement	Publication of personal content to boost engagement

.

Attracting more followers, sustaining engagement, and building the reputation and trustworthiness necessary for fundraising (Dowling, 2004) are all contingent on the successful actualization of Facebook’s affordances for CA: general affordances of visibility, association, and persistence, and specific affordances of association/interlinking and augmenting. While the affordances are available to all the profiles, profile owners can choose to actualise them to a lesser or fuller degree through the use of Facebook functions (Ronzhyn et al., 2022). In our case, transparency and authenticity represent different modes of affordance actualization—organisational actors leveraging persistence and association to communicate reliability, and individual actors using visibility and augmenting to project authenticity and sustain engagement.

Higher visibility of content can be achieved through a higher number of published content and more followers seeing the content (O1 and C1 having the broadest audience). A high ratio of reports to calls for action also contributes to the visibility as reports tend to be shared more: contributing to future donations as they appeal to a user’s self-efficacy. Evoking a positive emotional response from donation is necessary for donor retention (Wymer & Čačija, 2022). Conversely, Facebook’s approach to the visibility (and invisibility) of content coerces both groups into modifying their posting behaviour to comply with moderation and algorithmic preferences.

Association/interlinking affordance can help mitigate some of the differences in outreach. Using tagging, sharing, and hashtag functions, profiles can both increase the audience reach and improve their legitimacy and reputation. Tagging, for example, is used more often by smaller profiles (e.g., O2). Moreover, association/interlinking can aid the recognition of CA as a common cause (e.g., C1 and C2 reply to some of the comments by the users creating a feeling of community).

The persistence affordance of Facebook is crucial for building relationships with donors. It allows record and conserve the past results of the CA, one of the most important decision factors for the donors (Podolny, 1993). Persistent records also contribute to the transparency of organisational operations (Zhou & Ye, 2021). At the same time, the editability affordance allows profile owners to preserve the publication (along with its likes and shares), even if a part of it must be removed due to requests from the people in the photos or Facebook moderation requests.

Increasing engagement rates and post reach can be realised through the actualisation of the augmenting affordance: functions like photo and video uploads, live videos, and interviews can both increase engagement rates and influence donation behaviour. The use of emotions to invoke one’s empathy is a common way of stimulating donation behaviour (Wymer & Čačija, 2022). In this sense, engagement metrics function as observable indicators of how effectively each profile operationalizes the affordances of visibility, association, and augmenting to mobilise collective support.

This case contributes three insights to the study of collective action affordances on social media:

(1)

Different ownership structures lead to distinct affordance strategies. Organisational and celebrity profiles intentionally employ different affordances to achieve trust and participation outcomes: organisations rely on the intentional use of the persistence affordance to generate transparency and legitimacy, whereas individuals draw on the augmenting affordance to convey authenticity and emotional connection. These contrasting strategies illustrate how different actor types navigate the same platform affordances to foster engagement and mobilise collective support. The “hybrid” case (OC) demonstrates that these strategies can also be productively combined.

(2)

Engagement as a metric for affordance actualization. Interaction metrics (likes, shares, comments) are not secondary indicators, but evidence of how effectively actors realise visibility and augmenting affordances.

(3)

Platform governance shapes affordance use. Algorithmic moderation constrains the actualization of affordances, prompting strategic self-censorship and highlighting the negotiated nature of visibility in wartime context.

While our work is based on the analysis of content on Facebook, we believe that the findings can be generalised (albeit possibly with some reservations) to other social media as well. Other social media platforms also provide visibility, association, persistence, and other affordances discussed here. They may be realised to a different degree, and be more or less prominent compared to Facebook, but they are still present and can be actualised by CA actors. Differences in actions required for the actualisation of these affordances may be the reason behind the suitability of a particular platform for fundraising or other CA. As an example, Twitter’s hashtag function can make the actualisation of the association function easier, while Snapchat’s disappearing messages preclude the benefits granted by content persistence.

Beyond social media, there are clear differences between organisations and individuals involved in CA. Most notably, the need for organisational coordination for organisations. While Bennett and Segerberg (2012) suggest the absence of organisation coordination as the main peculiarity of social media CA, it is only true of bottom-up organising, not for CA arranged by formal organisations. In the case of fundraising, there is still a need for organisational coordination both for collecting the money and for putting it to use (acquiring equipment). Compared to individuals, in our sample, organisations had (1) access to equipment unavailable to individuals (for example, military use equipment), (2) the ability to acquire and distribute more and different types of equipment (both C1 and C2 focused on very few kinds of items, mostly cars and drones), (3) were better at redistributing the equipment as they had storage spaces and more personnel to work in logistics. Looking into the content of reports listing the acquired equipment and comparing it between the profiles, it becomes apparent that organisations are more apt for CA of this type than individuals (even if in social media, some individuals may have a much broader audience reach).

More generally, there are at least two unique aspects of the collective initiatives discussed in this paper in comparison to the previous research. Firstly, the wartime CA initiatives are not competitive among themselves. While CA in politics, business, or social activism often pursues mutually exclusive goals, in the case of the war effort in Ukraine, the general goal of all the collective initiatives is the same. It can be summarised as achieving peace (through the military victory of Ukraine) and minimising the war damage (both direct and collateral). In this regard, the activities organised by the various charity organisations and volunteers bear the features of a social movement: they are structured through dense informal networks, are geared towards developing and sustaining collective identity, and engage in CA to find solutions to identified problems (Della Porta & Diani, 2006). The second unique aspect is that while much of the previous research focused on non-monetary contributions to a political cause through sharing, liking, or following (Margetts et al., 2014), here the success of CA is contingent upon more significant actions, like donating money for the cause.

Finally, it is important to acknowledge the limitations of the research design. First, the sample is limited to the largest public fundraising profiles, so some effects that concern only smaller ones or personal profiles are not discussed. Similarly, there is no discussion of challenges that prevent profiles from becoming successful and attracting enough followers to manage fundraising successfully. Moreover, while the selected timeframe might be the most interesting in terms of profile development and growth, exploring a larger timeframe would uncover some new insights, particularly those connected with donation fatigue and new needs for the army as warfare progresses. Finally, as noted earlier, the editability affordance was not explored fully, as some edited and deleted posts were not included in the sample. While some of these limitations (i.e., timeframe) can be addressed in future papers, others are limitations imposed by the Facebook platform that are impossible to overcome using the same research methods (i.e., access to personal profiles is very limited).

8. Conclusions

This study explored how Facebook affordances were actualised by different actors engaged in collective action during the war in Ukraine. By examining both organisational and individual profiles, the study identified two distinct modes of using the platform for mobilisation. Profiles that made fuller use of Facebook’s affordances of visibility, association, and augmenting achieved substantially higher engagement rates, showing that user interaction can serve as a behavioural indicator of affordance actualisation.

The analysis revealed clear and consistent contrasts between organisational and individual (celebrity) profiles. Organisational profiles post heavily but see lower interaction per follower—typical of institutional communication emphasising transparency and persistence over emotional immediacy—while individual profiles publish more sporadic and less systematic reports, instead focusing on building a personal connection with their followers by providing regular updates on the life of profile owners and their activity. One profile (OC) that exhibited the properties of both an organisation (being legally a fund) and an individual profile (very much focused on the person for whom the fund is named), as might be expected, exhibited properties from both of the groups, demonstrating that these approaches are not mutually exclusive but can reinforce one another in practice. Both groups showed similar reliance on visual content and human-centred narratives: in the case of organisations, by describing the work and contributions of employees and volunteers, and in the case of individual profiles, by sharing personal connections to the profile owner and their relationships with donors and recipients of help. Both types of profiles highlighted the human dimension of their support: often showing people receiving help in their reports rather than simply the acquired items. The findings also highlight how algorithmic governance shapes the realisation of affordances, constraining visibility and prompting strategic self-censorship even within civic and humanitarian contexts.

Beyond its empirical insights, this study advances a systematic approach to collective action affordances in social media—addressing a gap identified in prior research. The present research contributes conceptually to the study of collective action affordances on social media. By interpreting engagement as a manifestation of affordance actualization, this study bridges the conceptual and empirical layers of social media-based collective action. This operational bridge between platform features and user outcomes contributes to affordance scholarship more broadly, offering a way to integrate quantitative engagement metrics into qualitative affordance analysis.

Future research could compare affordance use and engagement across multiple platforms or over longer timeframes to capture evolving patterns of mobilisation. Expanding the sample to include smaller or less visible initiatives may also reveal how affordance strategies vary with scale and resources, contributing to a fuller understanding of digital collective action during crises.