Two Steps Forward, One Step Back: The Evolution of Democratic Digital Innovations in Podemos

ABSTRACT While the digitalisation of political parties is increasingly analysed, less attention has been paid to the evolution of digital procedures and their consequences on intra-party democracy and party change. We propose a typology for identifying different types of evolution processes (consolidation, reconfiguration, mutation, and elimination) using the Spanish party Podemos paradigmatic case. Our analysis points out the centrality of hard and soft setbacks in the evolution of the digital procedures of the party. Findings indicate the relevance of different dynamics, such as institutionalisation, personalisation, and factionalism, jointly with other internal and external factors. Studying the evolution of digital party procedures is relevant for tracing party change in digital parties and other parties that are experiencing digitalisation processes.

party democracy procedures (García Lupato & Meloni 2021), among others. Podemos has also been conceptualised as a movement party (Della Porta et al. 2017), considering the modest investment in organisational structure, the lack of formal definitions of membership roles, the presence of a charismatic leader, and the idea of grassroots democracy. Even though the model considered may affect the analysis (Della Porta 2021), the characteristics of the digital party and the movement party are not necessarily exclusive but complementary. In this sense, studying the evolution of Podemos' digital-based processes and tools delves into party change in the digital and movement parties.
Adapting the discussion regarding democratic innovations (Smith 2009; Elstub & Escobar 2019), we can define parties' democratic digital innovations (DDIs) as new technology-based procedures developed by parties for increasing and deepening the members' and sympathisers' roles in the intra-party democracy. The DDIs may be successful but can struggle to consolidate, as with all innovations. These initial innovations developed by a party can evolve, and their objectives, implementation, and relevance may change. Some may prove inefficient or unable to fulfil their initial expectations and fail as a central part of the innovation process. On the other hand, innovations can hinder the electoral objectives of the party or limit the leadership's control. Due to their nature and regulation within the parties, DDIs may incur more significant changes than offline procedures.
There is a lack of literature on how some of the most salient parties' digital innovations have evolved. Contrary to previous party reform analyses, some changes may imply setbacks and retrenchment in the intra-party democracy and members' participation. This article aims to analyse party change in a digital party by studying the evolution of various digital procedures and exploring the relevance of setbacks, contrary to the idea of technology being an unequivocal mechanism for enabling more participation and democracy. We define them as changes in the aim, scope or technological implementation of internal processes and tools that imply a retrenchment or elimination of these initial innovations. This article focuses on the evolution of democratic digital innovations in digital and movement parties, mainly dealing with innovation setbacks. Hence, how do digital setbacks impact the evolution of new parties? What are the main factors that explain these setbacks? More generally, what can we learn by analysing setbacks for understanding party change?
In the next section, we will address the debate on party change and present our typology of the evolution of digital processes. Then, we will present our research framework based on the case study of Podemos. In the fourth section, we analyse seven procedures of the party and their evolution. The following section discusses the results and the last one concludes. of digital democracy (Gerbaudo 2019) and the limitation in the real impact of these innovations (Gomez & Ramiro 2019;Raniolo & Tarditi 2020).
The results are far from expectations. Following Barnea and Rahat's (2007) multilevel framework, digital (and movement) parties emerged in a political context that strongly criticised political parties' cartelisation and demanded more democracy at the systemic and party level. In a relatively short time, these parties successfully conveyed changes in the party systems and competition, making participation and democracy a critical defining feature of their policy proposals and internal organisation. Nevertheless, due to the increasing internal and external pressures (intensified by the institutional representation and even government status), these parties had to redefine, adapt, or reconsider certain internal procedures. As they were coming from open and participatory processes, some of their changes implied a retrenchment, limiting participation, modifying specific objectives or key aspects of the implementation, or directly eliminating certain procedures. Therefore, party change must also be analysed through the possible existence of distinct types of setbacks. Whether some procedures indeed became institutionalised, others experienced different evolutions. This process of party change is particularly evident in digital intra-party democracy procedures, as they are more subject to experimentation (Bennet, Segerberg & Knüpfer 2018, 1656. However, as parties are experiencing growing digitalisation and the tensions between equalisation and normalisation of the use of technology  3), digital setbacks may also be relevant for more traditional parties.
While some cyber-optimistic analyses claimed that new citizens' participatory demands and digital means 'might prove more positive for democratic engagement and the decentralisation of political power than has often been assumed' (Chadwick & Stromer-Galley 2016, p. 287), others considered that 'much of this technology-enabled innovation serves to reinforce the conventional model of vertical linkage' (Bennet, Segerberg & Knüpfer 2018, p. 1659. This article focuses on setbacks as possible outcomes of the party change.

A typology of the evolution of democratic digital innovations
For analysing the evolution of these procedures through time, we propose a descriptive typology (Collier, LaPorte & Seawright 2012) considering two specific dimensions: (a) Initial objectives: using the description or definition provided by the party, we focus on the initial motivation (Gauja 2017, p. 12) for implementing an initiative. It includes ideas, principles, and goals of the innovation according to the different dimensions of democracy. The initial objectives may be stable or variable over time.
(b) Implementation tools: we focus on the initial implementation tools and their evolution through time, emphasising digital aspects and the requirements for participating in each procedure. This analysis can show continuity (i.e. the implementation is stable or with just minor adaptations) or changes, such as whether the implementation tools are upgraded with new features or downgraded with reduced capabilities. We also stress the possible growing limitation for participating, reducing the possible number of participants or their options.
Thus, we have developed a typology for understanding the evolution of the parties' digital innovations over time (Table 1). We identify four types of DDI evolution: (1) Consolidation. This type applies to procedures that show continuity in their objectives and implementation tools. Some modifications, small technical changes or adaptations can occur, but most of the objectives and the procedure implementation are stable. The consolidation implies the institutionalisation of the procedure over time as a routinised way for the party to accomplish a specific internal objective. (2) Reconfiguration. This type applies to procedures that show continuity in the objectives but changes in the implementation tools to a meaningful degree. This process can develop in two different ways. On the one hand, it can take the form of an upgrade of the digital tools available for its implementation, implying a further development of digital innovation. On the other hand, changes can downgrade the available tools, indicating a retrenchment of the initiative. (3) Mutation. This type applies to procedures that change their objectives using the same implementation tools. Mutation can occur if the party redefines a procedure or its use over time. For example, the objectives may change since the initially expected consequences of a given procedure show different results than originally stated. Sometimes, the mutation process can imply a retrenchment or delimitation of specific procedures. (4) Elimination. This type applies to procedures that have changed their initial objectives and implementation tool(s). It can occur through cancellation, disappearing as a party process, or substitution by a different procedure with alternative objectives and/or implementation tool(s). Moreover, the initiative can be suspended or temporarily cancelled for its possible reconfiguration. It remains to be established how long a procedure must be suspended until it is considered effectively eliminated. Lastly, an initiative can be radically transformed if a party changes its objectives and invests in other technological solutions. Notably, creating a new digital process (and even renaming it) is more rational and efficient than profoundly changing the process and maintaining the name.
As we have pointed out, time plays a crucial role in the evolution of procedures. Since procedures can go through different processes, they should be categorised differently depending on a specific moment in time. For example, one procedure can start (T0), be technically upgraded (reconfiguration, T1), and then become consolidated or eliminated in T2.
The typology allows us to distinguish between procedures that have been consolidated and somehow routinised and others that have experienced changes. Among the latter, we point out the existence of setbacks, dividing them into two types. On the one hand, we define hard setbacks as changes implying that existing procedures are no longer available, which impacts the intra-party democracy since the party abandons the procedure to deliver a specific democratic goal (as in the cases of elimination). On the other hand, we refer to soft setbacks as changes that imply the retrenchment of processes and initiatives (as may occur both in the mutation and reconfiguration types). Those procedures are still active, but their initial objectives or implementation have changed, making them less participatory or inclusive. The typology and the notion of setback help to better understand democratic digital innovation and its impact on party change.
This observation is especially relevant if we consider new parties rapidly evolving in their structure and organisation due to their short life-cycle span (Pedersen 1982) and subsequent institutionalisation process (Randall & Svåsand 2002;. The latter is derived from their increasing role within institutions, including multilevel representation and governmental roles (in some cases). Consequently, as stated by Kitschelt about movement parties (Kitschelt 2006, p. 285), organisational learning could imply that 'critical elements of the original governance structure were abandoned'.

Case selection, framework, and methodology
This article is based on a single case study of Podemos, focusing on digital-driven processes and tools within its internal organisation for fostering intra-party democracy. As we have argued, Podemos is among the most recognised and studied digital and new movement parties. So, it can be considered a typical case (Seawright & Gerring 2008) or a paradigmatic one (Flyvberg 2006). As a challenger party (De Vries & Hobolt 2020), Podemos deployed strong anti-system claims and democratic deepening objectives, emphasising the quality of democracy and representation. Those claims and the protest movements' legacy have been mirrored in the party's objectives, aims and organisation, claiming to be more democratic, participatory, and transparent (Vittori 2017;Scarrow, Webb & Poguntke 2017). In this direction, technology and innovations have been identified as constitutive elements in their quest for internal democracy (Raniolo & Tarditi 2020).
The foundation of Podemos (March 2014) broke down the Spanish two-party system showing the difficulties of the Socialist and Popular Party at that time (Orriols & Cordero 2016;Rodríguez-Teruel, Barrio & Barberà 2016). Podemos' electoral performance improved during their first two years, passing from their first five members of the European parliament (MEPs) at the 2014 European elections to 71 members of parliament (MPs) at the 2016 general elections. In coalition with United Left and considering regional candidacies linked to them, they obtained 21.1 per cent of the votes, almost surpassing the PSOE (Partido Socialista Obrero Español -Spanish Socialist Workers' Party) as the main party on the left (falling short by just 1.5 per cent of votes). Since then, Podemos' electoral support has declined, losing almost 2.5 million votes and half of their MPs in the November 2019 general elections. Despite declining support, they joined the first-ever coalition government with the Socialists in January 2020 (Simón 2020), overcoming the electoral and relevance thresholds (Pedersen 1982).
Internally, Podemos has been characterised by difficulties in creating a strong and rooted organisation. The clash between different organisational proposals emerged during the first two party congresses (2014 and 2017). The growing factionalism, derived from contrasting ideologies and political strategies, led to party splits and the creation of new ones (in Andalucía and Madrid, led by former Podemos' founders). Finally, in 2021, Pablo Iglesias' charismatic leadership ended, calling for the fourth party congress in six years.
We can observe a mix of internal and external shocks that play a crucial role in understanding the party's organisational change (Harmel & Janda 1994;Rihoux 2016, pp. 303-305). These shocks and the institutionalisation push impacted the evolution of digital innovations within the party, especially when in government (Kitschelt 2006). These elements make analysing Podemos especially relevant for studying how party innovation processes and tools have evolved or adapted over time.
For this study, we have selected different DDI procedures developed by the party since its foundation. This selection considers their relation to specific dimensions of (internal party) democracy and the role of technology. They are different in scope, implementation, and impact. Importantly, their evolution intersects our typology, providing insightful cases for each process type ( Table 2).
All the procedures selected are somehow based on digital technologies for their implementation, from a simple webpage for registering to a platform for participating or an online voting system. Other possible initiatives that are not digitally based (such as party branches or their initial salary cap for all representatives) have not been directly considered in this study. Finally, the time frame is also relevant. As we are studying the evolution and possible changes in these tools and initiatives, we present their evolution through time, since its first approval and implementation (mostly around the end of 2014 and in 2015), the subsequent (possible) changes and adaptations and, finally, their actual form as of 2021. Considering Podemos' lifespan allows us to analyse how these procedures have adapted to the party's electoral results, passing from an extra-parliamentary party to a governmental one.
For the analysis, we rely on various sources. Firstly, we examined official party documents, including Party Statutes, organisational documents, and other internal regulations (such as rules concerning primaries or membership), and specific websites for certain initiatives. Secondly, we relied on publicly available data on the procedures' use, including voting and membership. For some insights, we also used newspaper coverage and the position of relevant party members. Finally, we carried out four semi-structured interviews with different members of Podemos. We interviewed one party representative at the municipal level, one technical officer at the national level, one former party representative at the national level, and one former central officer at the national level (see Online Annex, Table A).

The evolution of digital innovations in Podemos
Even if setbacks are the focus of this article, it is relevant to briefly analyse two procedures consolidated in Podemos: primaries and microcredits. Then, we focus on soft and hard setbacks experienced in five party procedures. We describe each initiative's objectives and implementation, tracing their evolution over time.

Primaries
Since its foundation, Podemos has always called open primary elections for selecting the elective offices within the party's organisation and the candidacies for representing the party in the elections 'from the first candidate on the list to the last' (Podemos 2014, Art.13). Primaries go in line with a process of democratisation at the political level (Barnea & Rahat 2007). This trend is especially true in Spain because after the first experiments during the 1990s, and since 2018, all relevant Spanish parties have organised primaries to select their leaders and/or candidates (Barberà & Rodríguez-Teruel 2021, p. 50).
Since the party's first Organisational Principles (Podemos 2014), most regulations on primaries to choose the candidates in the elections have remained unchanged in the subsequent Statutes (Podemos 2017a(Podemos , 2020a. Electronic voting is the default system, with some specific exceptions for the in-person vote (in smaller councils, as established in detail by Art. 3 of the 2018 primaries regulations, Podemos 2018).
Primaries for selecting internal party bodies follow a similar regulation, and they influence each other. In 2017, during the second party congress, the electoral system was changed by introducing a new system, named DesBorda, which has been used for all the subsequent party primaries. Arguably, this adaptation of the electoral system, which introduced a ranking point system for each candidate (and the possibility to select and vote for an entire list), has been the most controversial and relevant change in the regulations. For critics, this system tends to overrepresent the winning list to the detriment of the smaller ones, and its implementation is perceived as a betrayal of expectations regarding Podemos' openness and participation claims (Interviewee 1).
At the party level, we can observe that primaries in Podemos have been almost plebiscitary, benefiting party leadership personalisation in an otherwise fragmented party. Data from the six primaries for selecting the party leader or candidate show this lack of competitiveness since no candidate came close to challenging the mainstream one (the closest had a 78.2 per cent difference). It was true for the five primaries won by Iglesias and the last primaries during the fourth Congress confirmed this trend, since Ione Belarra won with 88.7 per cent of the votes (See Online Annexe Table B).
Podemos makes widespread use of these processes at all levels (and types of candidates), and it has been constitutionalised in Party statutes, showing a routinisation of this tool. Consequently, primaries have shown a high level of stability in their objectives and implementation and, therefore, can be considered a consolidated innovation within the party.

Microcredits
At the political system level, party financing is strictly regulated in Spain, and parties heavily rely on state subsidies, which amount to 70-90 per cent of their income (Rodríguez-Teruel & Casal Bértoa 2018;Ignazi & Fiorelli 2022). Additionally, Spanish parties are heavily in debt with banks at the party system level, and widespread corruption scandals have eroded citizens' trust. Accordingly, Podemos has also innovated in its financing to differentiate its organisation from other Spanish parties since the beginning.
In the specific case of elections, Podemos recurs to a digital financing innovation: microcredits. These are small civil loans from registered members (from €50 to a maximum of €10,000), with zero interest rates that, contrary to crowdfunding, the party returns once they receive the State's electoral subsidies according to their electoral result (around one year after the elections). The entire process for subscribing to the microcredit is conducted through a specific website that becomes activated when an election is called and is almost fully digitalised.
Podemos microcredits are part of the party's 'participatory financing' principles, which the party considers 'a necessary condition for the proper functioning of democracy' (Podemos n.d.). In particular, the party invented them to differentiate its organisation from indebted parties, asking citizens directly for money to support the party, its ideology, and policies. Consistently, the party's first ethical code stated the commitment to '[p]revent Podemos from participating in debt banking products' (Podemos 2017b), which then has also been included in the Art. 80 a) of the 2020 Statutes (Podemos 2020a).
This innovation has been stable since its creation in 2015 and has helped the party's institutionalisation by reinforcing its autonomy (Barberà & Barrio 2019, p. 266). Some minor changes have occurred (especially regarding technical and legal requirements), and the impossibility of bank financing has been incorporated into the Statutes. Nevertheless, the objectives and the implementation exhibited continuity through the years in a very regulated issue, fulfilling the party's objectives and is, therefore, a consolidated process.

Membership
Podemos membership has been one of the most cited new party digital membership cases. Its fluid nature (Deseriis & Vittori 2019) has been studied as an innovation to the traditional and multi-speed membership model (Gomez & Ramiro 2019). Gerbaudo (2019, pp. 17-18) defined it as an 'open membership model' similar to the registration on social media such as Facebook with 'minimal membership requirements' (Gomez & Ramiro 2019, p. 536). Initially, unlike the multi-speed membership model (Scarrow 2015), Podemos offered just one form of affiliation that enabled participation in all the party's activities and internal decision-making processes (Gomez & Ramiro 2019, p. 537). While traditional parties aimed to increase and strengthen membership (Gauja 2017, p. 32), Podemos was characterised by a very inclusive and participatory model from the beginning.
In the Organisational Principles document (Podemos 2014, Art. 10), full participation rights were recognised to 'all the people who are part of Podemos' as members of the Citizens' Assembly (permanent party's body). This regulation challenged the distinction between inside and outside the party, consistent with Kitschelt's movement party (Kitschelt 2006, p. 280). Three years later, the Podemos (2017a, chapter II) regulated membership in detail and included three overlapping categories (see Table 3). Additionally, a verification process was required by uploading an identification document to certify the member's identity. Finally, in the new Podemos (2020a), approved in the third congress, the categories were modified, and a membership quota was introduced.
The evolution that occurred in 2020 significantly changed the membership model, bringing it closer to a multi-speed model that differentiates between full members who financially contribute to the party periodically, and other types of members with lower commitment (Scarrow 2015). Therefore, one of the two main characteristics of this innovation disappeared but, through the same tool, Podemos is pursuing somewhat different objectives. Although no party decisionmaking process has been limited to full members, the statutes' changes marked the choice of moving from a fluid model to a differentiated and demarcated membership with a financial link with the party (currently established at €3 per month). The evolution of the membership towards a traditional affiliation could be considered a marker of institutionalisation, and it may provide the party with greater opportunities for control. Accordingly, the party's census passed from more than 500,000 members to 138,847 verified members (Europa Press 2021) and 18,791 full members (Bocanegra 2020) in 2021 (see Online Annexe, Table C). Hence, we can observe changes in the party's membership promoted by altering the formal rules (Gauja 2017), which constitute a mutation from the initial idea of fluid membership to the actual multi-speed one. We define this process as a soft setback since the membership's evolution has implied a distinction between distinct types of members and their role in decision-making, somehow experiencing a retrenchment if compared to the initial movement party model.

Consultas ciudadanas
The Consultas Ciudadanas (Citizens' Consultations) are binding consultations 'on matters of special political relevance' (Podemos 2020a, Art. 15) voted online through the Participa platform by the Citizens' Assembly. The online platform potentially allows fast, frequent, and cost-reduced use (Deseriis & Vittori 2019). It offers a good example of using technology as a democratising tool, allowing members to position themselves on relevant and timely issues. Since its development, the Consultas have been a crucial feature of the Podemos decisionmaking process, thus, reflecting a participatory mechanism that distinguished the party from its competitors. During the various congresses and in the party's documents in force over the years, this tool has essentially maintained its main formal characteristics, with some exceptions (see Online Annexe, Table D). Notably, the link between this participation tool and the deliberative one (Plaza Podemos) disappeared, as well as the possibility for members to propose, debate and vote for particular policies or political positions. On the other hand, the evolution of the criteria for member-based consultations became more stringent, and party branches cannot call them anymore. 1 It is plausible that these changes were in response to a centralisation push towards institutionalisation, with growing power in the state-level organisation (especially the Executive branch) and a lack of internal (online) debate on the issues at stake. Consistent with Gerbaudo (2019), the usage of this procedure may have evolved into a more plebiscitary tool, reinforcing the leadership's position. From 2014-2021, Podemos organised 12 consultations at the national level (see Table 4). Each consultation supported the leadership's position with an overwhelming mean support of 87.7 per cent. All of them have been called by the leader or Citizens' Council, and none by the members or branches. Hence, while the initial objectives could be linked to bottom-up participation, this procedure has evolved towards a leadership-centred and controlled procedure.
Two cases are especially relevant for showing the mutation of this procedure. They point out the importance of personalisation dynamics and the role of internal factions. Personalisation and leadership's use of the procedure was evident in May 2018, when a contested revocatory consultation was called on Pablo Iglesias and Irene Montero after purchasing a €600,000 cottage (El Mundo 2018). The same secretary general decided the timing and question and led to a plebiscite on the party leadership, with increasing accusations of appropriation of the participatory tool for 'legitimising the whim of the leadership' (El HuffPost 2018).
The second consultation was on the investiture vote of PSOE's leader Pedro Sánchez as Prime Minister in July 2019 (which finally failed and ended with new elections in November). This strategic decision was divisive within the party.
There was criticism about the timing (in the middle of the negotiations) and the options offered, which did not reflect all possible scenarios. Teresa Rodríguez, the secretary general in Andalucía in that time and one of the leaders of a critical internal faction, claimed that the options were 'overtly tendentious', did not include other possibilities, and that the Consulta was 'regrettably, a real insult to the intelligence' (Rodríguez 2019).
Finally, it is worth noting that since Podemos joined the government (in January 2020), no Consultas have been promoted, despite the disagreements between PSOE and Podemos within the coalition on relevant issues. The government threshold may limit the possibility of the Consultas, and the consequent step forward in institutionalisation produced a mutation in the process. While formally maintaining its objectives and implementation, the procedure moved from a participatory tool to a more reactive, centralised, or plebiscitary one (Gerbaudo 2019), supposing a soft setback in the initiatives' evolution.

Plaza Podemos
Plaza Podemos (Podemos Square) was Podemos digital deliberative tool from 2014 until July 2019. In the Organisational Principles document (Podemos 2014, p. 7), it is presented as 'a space for debate and deliberation' for collective decision-making and for creating 'ideas, projects and Notes: * The data presented is a total average of the votes on all programme points. **This consultation included two different questions; we selected the second one relating to Podemos. ***The numbers preceded by the sign ≈ refer to approximate values. Source: Authors' own elaboration based on primary sources (i.e. party website) and secondary sources (e.g. newspapers coverage) for missing consultations, due to the absence of publication of a voting archive.
proposals'. Annexe I of the same document stated the 'totally free and open' nature of the tool ('anyone can easily propose, from inside or outside Podemos') and the close connection with the Citizens' Assembly of the party, especially with the Consultas Ciudadanas. Through Plaza Podemos, it was possible to upload and debate on proposals called Podemos Citizen Initiatives (ICP, Spanish acronym), linking the principles of deliberative and direct democracy to further democratic innovations (Smith 2009, p. 11). Participants could propose and endorse different proposals in a three-step process (see Table 5). While the objectives remained similar, on 31 October 2015, Plaza Podemos shifted from Reddit (social news website and forum) to the Consul application (open software platform), inaugurating Plaza Podemos 2.0, thus moving from an open network to an internal digital platform. 2 The justifications for the choice have been mainly technical, focusing on the opportunity for an internal platform customised to the party's needs (Interviewee 2). This shift went towards the institutionalisation and centralisation of the party's tools. It allowed for greater control and protected the leadership from the growing criticism on Reddit. Moreover, it introduced another mechanism called Escaño Abierto (Open seat) that permitted questions to Podemos' representatives and parliamentary groups, promoting vertical interactions rather than deliberative purposes.
Plaza Podemos 2.0 main problems were related to its effectiveness in participation and decision-making power. For example, between October 2015 and October 2016, the number of citizens' initiatives dropped from 1,405 to 407 (2018), and the mean votes per proposal fell from 198.3 to 17.6. 3 Furthermore, within the Podemos' deliberative digital tools, no ICP or other bottom-up vote proposals (as in the Consultas Ciudadanas) have ever reached the voting phase. Thus, the digital deliberation platform did not achieve the expected results, in line with the mixed evidence on the capacity to produce meaningful discussions through ICT-driven deliberation (Ivernizzi-Accetti & Wolkenstein 2017, p. 104).
After the first Podemos congress, Plaza Podemos disappeared from the main party's documents, and in July 2019, the Territorial Support Office replaced it. Thus, the party responded to the inefficacy of the deliberative procedure by developing a centralised tool with different objectives consistent with information purposes rather than deliberation. Accordingly, it is a one-stop shop for asking questions to the party via three top-down communication channels: email, phone, and an online form. It claims to offer all organisation members help or advice on organisational, discursive, procedural, or financial issues (Podemos 2019).
Consequently, Plaza Podemos displayed an interesting evolution through time. It started as an open, deliberative platform, was reconfigured with  implementation changes, and eventually eliminated and substituted by a non-deliberative tool. It provides a paradigmatic case of a hard setback.

Banco de Talentos
The main objective of the BdT (Banco de Talentos -Bank of Talents) tool was to promote the growth of the party through open and merit-based participation 'without going through exclusion mechanisms' (Podemos 2014, p. 6), incorporating the know-how and experience of volunteers, sympathisers, and society in general. It was implemented through a digital tool where interested citizens could complete and upload an online inscription form with their personal information, availability, and interests. The BdT initiative aimed to open the party to society by incorporating new profiles, providing a repository to the party that could be accessed for identifying the most suitable individuals (voluntary or paid) for supporting initiatives, campaigns, or structures. In the noninstitutionalised stage of the party, it was linked with the first phase of the foundation and validation of the party branches (see Plaza Podemos 2014). The BdT was managed directly by the Participation Area and is mentioned in the Organisational Principles document (Podemos 2014) approved during its first Congress, but it did not appear in any primary party documents since then. However, it is included in other documents of that Area, such as the calls of the three editions of Impulsa (see below), where the BdT is offered as non-economic support to project proposals (Podemos 2015b(Podemos , 2015c(Podemos , 2016b. Despite being presented as one of the party's most essential open tools in the first phase, the BdT disappeared without Podemos ever systematically using it. Indeed, eliminating this innovation has primarily been a consequence of its non-use. Unlike other innovations analysed, the BdT has not experienced changes in its objectives or implementation. Instead, it seems to have encountered internal party dynamics, where loyalty has prevailed over meritocracy, which this initiative aimed to promote. At the local, regional, and national levels, the selection processes of the party's collaborators (for the various positions) did not use the BdT, with occasional exceptions in the first phase. The selectors tended to appoint personnel through 'trusted people' (Interviewee 3), who used to be close to the same faction, reinforcing the party's fragmentation.
Podemos eventually eliminated the BdT without any aftermath, and at the moment of our analysis, there are only a few traces of it on the web. This innovative tool was developed by a party open to society and internal meritocracy but failed to achieve an equally innovative use in the party's selection processes. Thus, it offers another example of a hard setback.

Impulsa
Impulsa (Impulse) has been a Podemos initiative designed to support 'innovative projects with social projection' (Podemos 2016a), presented by people or non-profit organisations (not necessarily linked to the party), and chosen by registered members through a participatory process (Podemos 2015a). Impulsa was developed and regulated by Podemos' Participatory Area and was not included in any party document approved in Congress. Its first edition was launched in April 2015 and was closely related to the salary cap of Podemos' representatives. 4 The surplus funds were invested in Impulsa for financing projects within and beyond the party's activities/structure. The initial objective of this programme was to blur the boundaries between the party and civil society, in line with a movement party, in particular, linking Podemos with its social movements base. The launch of this initiative aimed to give a clear sign of a democratic and social distinction compared to traditional parties and has evolved during its three editions (see Online Annexe, Table E).
Associations and party branches could submit project proposals in the three editions, and the members registered on the Podemos platform voted on them. Until 2016, Impulsa was one of the party's most actively participated initiatives, which according to Miguel Ardanuy, coordinator of the Participation Area at that time, manifested 'the essence of Podemos' (Podemos 2017c).
However, the gradual reduction in proposals and voting possibilities over time mirrored a separation within and outside the party that Impulsa initially aimed to challenge. Similarly, the introduction of Hacemos (We make), a category aimed at financing internal party projects, exhibited a shift away from the movement party approach towards the internal structure.
After the third edition (January 2017), Impulsa was suspended. It responded to internal and external dynamics. On the one hand, the suspension marked a modification in the party's will since it claimed to orient its investment towards internal structuring. It has been another driver of party institutionalisation, pointing out the limits and tensions of a movement party, as stated by Kitschelt (2006). On the other hand, the salary cap (which surplus funded Impulsa) became a very salient issue and a source of media and other parties' attacks against Podemos due to some inconsistencies in its application. In a certain way, this initiative backfired on the party, resulting in the loss of project funding.
The suspension has been criticised, pointing out the decline of the relationship between Podemos and social movements and organisations, which risks distancing Podemos from some of its founding principles (Interviewee 4). Since the Impulsa programme evolved from its initial implementation, it was reconfigured, suspended, and, after five years, we could argue, eliminated; it thus provides an example of a hard setback.

Party change and the digital dimension: which lessons?
By analysing different procedures in Podemos, we got two important lessons. Firstly, the study of party change should focus on the main functions of parties (leadership/candidate selection, membership, and policy development, Gauja 2017), but also include other procedures that can help to grasp substantial changes in different dimensions of the intra-party democracy, especially if considering the digital dimension. Secondly, party change should be understood as a process, and analysed in its evolution over time.
Applying our typology reveals that while some innovations have been consolidated and relatively stable, others have mutated, been reconfigured, or even eliminated. In general terms, the evolution of the procedures points out a 'tension between control and interactivity' (Chadwick & Stromer-Galley 2016) that characterised Podemos as a digital and movement party along different processes (such as democratisation, institutionalisation, personalisation, and fragmentation). The changes observed had relevant consequences for internal party organisation and led to soft and hard setbacks. In Table 6, we summarise the evolution of each procedure, considering three moments in time, and highlighting the main driving factors.
Our typology shows different evolution patterns depending on the relevance of the process, its codification, the equivalence to other party change trends, and the dimension of democracy involved. Consolidated procedures are relevant for party survival. They are codified in the party's Statutes, facilitating certain stability, and the incentives for change are aligned at all levels (political, party system and party), which make their effect longer lasting (Gauja 2017, p. 12). Mutated and consolidated procedures share similarities, particularly in terms of codification. Nevertheless, mutated procedures experienced soft setbacks and certain retrenchments in objectives or implementation. However, in their current form, they converge to democratising reforms other parties have adopted to gain legitimation. Finally, when the procedures were eliminated, they were not part of the party's Statutes. Hence, experimentation, innovation, reconfiguration and elimination have been easier for the party, and they generally refer to specific dimensions (such as deliberation and inclusion) that are considered less relevant for party survival. However, they provide different insights that help better understand the dynamics of party change where the notion of setback, we argue, is crucial. Instead of promoting democratisation, increasing participation or the role of members in decision-making, party change may also imply a retrenchment or elimination of certain democratic processes, moving towards plebiscitary processes, leadership control, centralisation, and personalisation. Digital intra-party democracy procedures are especially suitable for analysing setbacks, but the same analysis can also be applied to offline procedures.
Interestingly, the setbacks were not due to technological problems or technical limitations 5 but instead caused by internal party decisions. This observation is especially relevant if we consider the debate between cyber optimists and more realistic accounts (Chadwick & Stromer-Galley 2016;Bennet, Segerberg & Knüpfer 2018). Our analysis identified other factors that play a central role in party change, effectively limiting the possibilities of digital innovations in parties. Understanding why this differential evolution occurred is highly relevant for comprehending party change, the digital parties' institutionalisation, and the effects of technology and innovation on political parties. Although further research is needed, our analysis provides five possible explanations and factors.
First, the tensions between an initial organisation as a movement party and the institutionalisation process could contribute to this evolution. Podemos institutionalisation is complex, especially regarding its multilevel setting (see . The setbacks of specific procedures that aimed to blur the distinction between the party and civil society (such as membership, Impulsa or BdT) are especially relevant. It shows the abandonment of one of the notable features of a movement party along with its institutionalisation. These changes align with internal shocks, the reconfiguration of the dominant coalition, various party splits and problems with electoral coalitions in different regions (Harmel & Janda 1994), and the ideas of personalisation and the role of leadership (Barnea & Rahat 2007).
Second, centralisation plays an important role in party evolution. Notably, it has been one of the core conflicts within Podemos, exemplified by the tensions between the two most charismatic figures of the party, Pablo Iglesias and Íñigo Errejón, and the reconfiguration of the dominant coalition (together with growing tensions with the Anticapitalist faction). Since Iglesias won the second congress (and some months later, Errejón left the party and created a new one), vertical integration has been high (Barberà & Barrio 2019, p. 258), and different conflicts with regional coalitions have surfaced (e.g. in Galicia, Catalonia, and Andalusia). These internal shocks produced incentives for party change that have evolved towards centralisation and personalisation, as is reflected in the evolution of Citizens' Consultations. As we argued, the use of the Consultas mutated, moving from a participatory mechanism to a more strategic and leader-oriented tool. In a similar albeit different way, Podemos' landmark deliberative tool, Plaza Podemos, was eliminated and replaced by a top-down initiative. In this sense, centralisation and disconnection between deliberation and decision-making seem to play critical roles in certain setbacks.
Third, internal and external dynamics are also relevant. In the case of the BdT, loyalty to a faction was preferred to meritocracy, showing an internal dynamic hardly surprising. Thus, a well-known internal party dynamic could drive the setback in innovation, but external dynamics are also relevant, as shown in the case of Impulsa. Even if other factors may be relevant, dynamics such as mediatisation, competition, and the party's image at the party system level must also be considered.
Fourth, successful innovations are part of the value-infusion of the party but are also aligned with the three dimensions of party change. Online primaries are part of Podemos' genetic model as institutionalised democratisation processes. Moreover, they follow a European and Spanish trend of primaries contagion (Sandri & Seddone 2021). However, they have been hardly competitive and characterised by personalisation, at least in the different leadership primaries (at the national level). Microcredits are also embedded in the principles and values of Podemos (in its Ethical code and Statute), reinforcing the party's identity and reputation with respect to other parties, denouncing corruption scandals and bank debt. Both these innovations have worked well, favouring routinisation and autonomy in the party's institutionalisation (Barberà & Barrio 2019, p. 266), and are therefore consolidated within the party.
Finally, some of the key characteristics of the digital party seem to have been retrenched. Accordingly, we should consider the evolution of certain digital procedures for analysing these types of parties. Regarding movement parties, Kitschelt (2006) argued that critical elements of the original governance structure could be abandoned while others stay, as in the case of the Greens (Rihoux 2016). Furthermore, Podemos' case offers an interesting intersection between digital and movement parties. Following Della Porta (2021, 1345), the latter implies a 'complex system of communication' with various online and offline technologies used by multiple players, in different arenas, with diverse dilemmas, and adaptation through trial-and-error.
Focusing on digital procedures only tells one part of the story, albeit an increasingly relevant one. As the setbacks are not due to technical problems or limitations but instead to internal and external processes of the party, they provide a good proxy for observing the evolution of movement party characteristics. In the debate about innovation and normalisation (Jacobs & Spierings 2016, pp. 27-34), we found that new digital parties are de-digitalising certain procedures, while other established parties are increasingly digitalising, thus representing a convergence in the digitalisation process (García Lupato & Meloni 2021).

Conclusions
Research on the digitalisation of political parties is gaining increasing relevance. Digital parties have implemented new processes that have received academic attention, demonstrating different innovations for intra-party democracy, its uses, and mixed impacts. However, less attention has been paid to the evolution of these innovations.
Through the paradigmatic single-case study of Podemos, we have analysed seven digital innovations and their evolution by proposing and testing a typology. Contrary to studies that equate party change and democratisation, we highlight the importance of setbacks as a possible outcome of the process of change. While some procedures were consolidated (e.g. microcredits and primaries), other well-known ones experienced hard setbacks such as elimination (Plaza Podemos, BdT, and Impulsa) or soft setbacks such as mutation (Citizens' Consultations and membership).
In the context of growing party digitalisation, our analysis is important for understanding the evolution of digital parties and movements, but it can also be applied to the parties that are digitalising their organisation. Therefore, when analysing party change, on the one hand, the digital dimension must be considered because digital procedures are more open to experimentation and can provide insights into parties' attempts to change. On the other hand, to analyse the evolution in time of different types of procedures from different dimensions of intra-party democracy is needed. Indeed, they both can provide relevant insights into party change, the effects (positive or negative) on intra-party democracy, and the logic of digitalisation in a broader range of political parties.