Assessing Inclusivity Through Job Quality in Digital Plat‐Firms

A great deal of the literature has underlined how job quality is a key element in individual well‐being. However, the rise in platform work challenges this issue, since not only do “plat‐firms” play an increasingly important role in job matching, work organization, and industrial relations, but they also increase the risks of a poorly inclusive socio‐technical system in terms of the quality of working conditions and accessibility. In this sense, the platform economy is intertwined with multiple forms of social exclusion by acting on pre‐existing inequalities that stratify workers within the labor market. This is particularly true in Italy, a country with a strongly dualistic labor market, which leads to a remarkable gap between insider and outsider workers. Therefore, the goal of our analysis is to evaluate the impact of the platform model on job quality in the Italian context. This will be accomplished by adopting an integrated and multidimensional perspective through the application of the OECD Job Quality Framework. The analysis identifies how job quality is differently affected by the type of platform work involved in terms of creating differentiated patterns of social inclusion/exclusion in the case of platform workers.


Introduction: The Digital Inequality Stack and Job Quality in Platform Work
In the prodromal phase of digital transformation, the internet was seen as a tool for social inclusion capable of changing the current economic paradigms through an economy based on information redundancy, accessibility, and community participation (Barbrook & Cameron, 1996).This rhetoric was nourished, on the one hand, by Marshall McLuhan's theories on the emancipatory power of the media and, on the other, by the liberal ideal of "catallaxy" (Hayek, 1978), according to which the internet would guarantee a meritocratic and more efficient system through disintermediation and the absence of a centralized control agency.This techno-solutionist and naive vision of technology was then overcome, not only by the transformation of the internet itself into a corporate platform complex (Terranova, 2022), or rather a socio-technical financing and infrastructuring process that concentrated the power of the web in the hands of a few actors, but also by the huge amount of critical reflection on the limits of digital technology and on the relevance of a plurality of digital divides that make the internet a reproducer of old inequalities and a generator of new ones (Castells, 2001).
Nowadays, platforms play an increasingly important role in job matching (Kässi & Lehdonvirta, 2018), work organization (Huws, 2017), and industrial relations (Duggan et al., 2020).However, digital transformation in the labor market generates new inequalities that intersect and stratify with older ones, building a "digital inequality stack" where multiple layers of foundational imbalances overlap and accumulate (Robinson et al., 2020).It is widely observed how platforms and digitalization reproduce and amplify existing inequalities (gender, class, racial, spatial, etc.) through the persistence of some access gaps and matching algorithmic dynamics that favor highly homophile transactional networks (Edelman & Luca, 2014;Tubaro et al., 2022).Even during the pandemic, platforms took advantage of the crisis by moving around "grey areas," further entrenching precarious and informal forms of work (Howson et al., 2022).
In this sense, platform economy emerges as being at the crossroads of multiple sources of social exclusion in the labor market, strengthening inequalities between insiders and outsiders (Huws, 2017), formal and informal workers (Farinella & Arcidiacono, 2023), experts and amateurs (Cingolani, 2021), supplemental and dependent earners (Schor, 2020), and paid and unpaid workers (Casilli, 2017).Such a role deserves special attention in Italy, a country characterized by a dualistic labor market, which creates a deep cleavage between an area of "core" jobs, with good working conditions and desirable material and immaterial rewards, and an area of peripheral bad jobs, characterized by poor conditions and low job quality (Scherer, 2004).This division further problematizes the already existent inequalities, like those on a gender, age, and education basis, and magnifies the role of platform work as a potential driver of inclusivity in the labor market (Cirillo et al., 2023).

The Quality of the Platform Work
Since the late 1960s, a great deal of the literature has explored the concept of the quality of work as a key element of individual well-being (Piccitto, 2022), which in the last few years has been challenged by the spread of platform work.The concept of job quality has been approached in different ways, depending on the scientific field under consideration.Generally, within the field of economics, job quality has been proxied by means of hetero-directed extrinsic job characteristics related to the system of rewards such as pay, job security, and fringe benefits (Howell & Kalleberg, 2019).Studies in the field of psychology, instead, are more focused on the internal worker's individual experience and the extent to which their psychological needs are fulfilled during the working experience (Piccitto, 2022).Finally, sociologists are more interested in defining job quality in terms of skills and autonomy (Gallie, 2012).Currently, scholars have provided evidence of the importance of an integrated and multidimensional conceptualization of job quality (Muñoz de Bustillo et al., 2011;Oesch & Piccitto, 2019) as the most effective way to view such a phenomenon.
In recent years, the unfolding of the digital economy has triggered a lively debate on the impact of digitalization on job quality.In this regard, commentators have polarized around two general visions: First, the "optimistic" view underlines how the digital economy can increase work flexibility from a post-Fordist perspective, increasing the worker's autonomy in carrying out work and improving the chances of an acceptable work-life balance (Mulcahy, 2017); moreover, it is argued that online platforms can facilitate access to work for traditionallydisadvantaged segments of the workforce (young people, immigrants, people living in inner areas; see De Stefano, 2016).Secondly, the more "pessimistic" view, which underlines, instead, how the digitalization of production activities accentuates the fragmentation of the work process, tracing the practices of the Fordist organization in a context of high contractual uncertainty (Healy et al., 2017), thus opening the doors to a "race to the bottom" in terms of wages and working conditions.
Several factors associated with the concept of job quality, both extrinsic (i.e., referring to the most basic and concrete aspects of work) and intrinsic (more emotional and less tangible), are being challenged by the impact of the platform economy.Concerning extrinsic factors, one of these is the formal definition of the worker's status, a feature that is increasingly ambiguous and difficult to define in light of increasingly elaborate and technology-driven models of new ways of organizing work (Healy et al., 2017).The lack of formal recognition of new digital workers is reflected in proposals for the creation of new "legal categories" by which online workers can identify themselves (Todolí-Signes, 2017).This ambiguous and intrinsically non-standard way of regulating digital working relationships translates into precariousness, which in turn leads to the exclusion of the worker from social protection (Donovan et al., 2016;Kalleberg, 2012;Schor et al., 2020) and peculiar forms of collective action intentions (Politi et al., 2022).
Anyway, there are not so many analyses that interrelate inclusion, job quality, and platform work.Traditionally, the analyses available on platform work have focused on specific categories of digital workers (e.g., riders, Uber drivers, etc.) or take into consideration specific dimensions of job quality (salaries, contractual status, access to social protection schemes, etc.), without adopting an integrated and multidimensional perspective (Behrendt et al., 2019;Berg, 2016;Fabo et al., 2017;Wood et al., 2019).
De Groen and Maselli (2016) pointed out how crucial it is to consider the plurality of platform work: Most jobs in graphic design or IT consultancy could be performed virtually, while others, like care or delivery/mobility services, need specific locations or physical interactions.Moreover, the level of required skills within platform work is heterogeneous: Many platforms are oriented to low-or medium-skilled tasks such as data entry activities typical of microwork platforms, cooking in social eating platforms, or writing and/or translating small amounts of text for e-commerce platforms; conversely, other platforms specialize in high-skilled professional services such as those in the legal or architectural fields.Hence, De Groen and Maselli (2016) identify two levels of skills associated with digital jobs, distinguishing between jobs that require high levels of human capital and educational credentials and those that involve performing basic tasks that require limited expertise, often without any specific formal credentials.Consequently, the two scholars suggest classifying platform workers into four categories that derive from the combination of these two variables (virtual/in-presence tasks and low-medium/high skill required).Such differentiations play an important role when assessing job quality and inclusivity in the platform working environment: For example, some studies have shown that physical/local tasks are comparatively better rewarded due to a smaller pool of workers (Aloisi, 2016).In those platforms providing virtual services, supply and demand are unbounded, without any limits, and they can be very quickly matched, overcoming spatial limitations (De Stefano, 2016).This issue appears to be particularly relevant for workers living in inner urban areas (Greene & Mamic, 2015) and for those of the Global South (Aleksynska et al., 2019), who may lack better job opportunities in their local context.On this issue, a growing literature recognizes how there is a digital aim to configure a more global and "planetary labor market," while, at the same time, the quality of work changes with different local contexts and platforms (Graham & Anwar, 2019).The peculiarity of platforms lies precisely in the ambiguity of their effects on the global market.For example, in some cases, they make employment relationships more visible and recognizable, especially in the Global South, where they represent a concrete opportunity for work and for the formalization of labor relationships; on the other hand, especially in the Global North, platform work is seen more as a transitional and complementary job based on long-rooted national strategies of labor market deregulation (Weber et al., 2021).
Concerning other job characteristics, platform work is characterized by a degree of personal control and flexibility that make workers agentic in selecting duties to be done, setting their own schedules and pace, and negotiating rates (Teodoro et al., 2014).Furthermore, such workers, intermediating virtual work, have the chance to work from home, a characteristic increasingly at the core of the debate on job quality, especially in post-pandemic times (Eurofound, 2020).These characteristics impact work-life balance chances (Rodríguez-Modroño & López-Igual, 2021) and promote female participation in the labor market (Chung & van der Horst, 2018).At the same time, however, these features can have fewer desirable side effects: They can lead to the intensification and extension of working time and an overlap with the worker's sphere outside of work; this dynamic is known as the "autonomy paradox" (Mazmanian et al., 2013) and is particularly harmful, especially for women.Additionally, individuals working on platforms are at risk of social and professional isolation, because of the nature, location, and organization of their work (Durward et al., 2016).
In conclusion, we have outlined how current research looks specifically at single types of platforms and rarely with a comparative perspective.Consequently, the platform heterogeneity in terms of job quality and social inclusion has been largely underexplored.

Objectives and Methods
The present article explores the issue of job quality in Italy's platform economy (for a similar approach see also Arcidiacono et al., 2021), trying to focus particularly on the extent to which digital work succeeds (or not) in improving the workers' conditions, especially for those operating in more peripheral and disadvantaged sectors of the labor market.In this sense, our analysis links the issue of social inclusion to that of the quality of work in the platform economy, with a comparative approach that considers the heterogeneity of performed tasks, work engagement, and working conditions (Howcroft & Bergvall-Kåreborn, 2019;Kalleberg & Dunn, 2016).Furthermore, we explicitly recognize the multidimensionality of the concept of job quality (Muñoz de Bustillo et al., 2011;Oesch & Piccitto, 2019) and that different concepts and characteristics tap into the domain of job quality (Steffgen et al., 2020).In this view, our framework is inherently cross-disciplinary and integrated, contrary to the approach that is generally adopted in the literature (Findlay et al., 2013).In particular, we decided to adopt the OECD Job Quality Framework (Cazes et al., 2015; see also Figure 1).This framework was developed based on a careful recognition of the various indicators and dimensions used by the international community to evaluate work quality and make comparative analyses between different socio-economic groups.The choice was inspired by the coherence of this approach with the available empirical material and is characterized by a focus on results (outcomes) as well as on workers' subjective voices (Dunn, 2020;Frenkel, 2015), rather than on the drivers of the quality of work per se.Starting from the areas of well-being identified by Stiglitz et al. (2009), we define job quality in terms of the three axes presented in Figure 1.
Our analysis also considers heterogeneity across platform types and the workers who are active in them.In particular, we decided to adopt the classification of platform jobs proposed by De Groen and Maselli (2016), distinguishing between tasks that are electronically transmitted (virtual/global) and those that require manual labor or physical interaction (physical/local), and between low/medium-skilled and high-skilled jobs (see Figure 2).
We select 20 platforms that represent the different types of platform jobs.The selection is the result of a preliminary activity involving digital mapping of the platforms active in the Italian context through a systematic search on Google.The first results of these mapping activities are presented and combined with the results of consultation with key informants within two exploratory focus groups.These focus groups included workers, experts, and exponents of business interest associations, and were also important in terms of identifying issues and questions for the subsequent phases of our analysis, especially concerning platform selection.
The 20 platforms were selected based on the number of subscribers they had (on the demand as well as the supply side) and the platform's presence in the reference market (meaning the regions where the platform is active).Combining these criteria with indications by the focus group experts, we identified the most relevant platforms in terms of the national labor market.Our methodology employed different data collection techniques, namely: 1.A net-ethnographic analysis (Arcidiacono, 2019;Kozinets, 2009) on the organizational design of the selected platforms, using an observational diary organized in terms of (a) job matching mechanisms, (b) reward systems, (c) relationships/ community, and (d) user experiences.2. 41 semi-structured interviews with managers and founders (13) and platform workers (28) from the 20 platforms selected in the previous phase.
The mechanism to select and identify the interviewees involved a snowball system, through direct contacts on the platform or related social media profiles.In particular, our interviewed workers were recruited online, on Facebook groups, or directly on the platform where they worked.

Earnings Quality
An analysis of the quality of the earnings in platform work considers not only the amount the workers make but also how it affects their material well-being.work with casual or informal employment, and so they evaluate platform earnings as being significantly better.A manager at Helpling Italia explicitly stated that undeclared work is the real benchmark in their business model.Hence, their employees' favorable evaluations of earning quality come from the fact that low-skilled workers, aware of their marginalization in the labor market, will compare their objectively undesirable online jobs with the even less desirable conditions of informal or undeclared work, expressing a "satisficing" attitude toward their situation (Walters, 2005).On the other hand, in the case of low/medium-skilled digital work, workers declare that they obtain on average lower wages when compared to offline low-skilled jobs.The same happens to high-skilled work (designers, architects, lawyers, etc.), where the reduction of costs for clients implies an increase in competition between professionals (Arcidiacono et al., 2023) and a deterioration in pay conditions.However, the assessment of earning quality depends strictly on workers' employment status.In most cases, earnings on the platform are considered as a complementary income to that obtained from a main offline job.This element highlights a first important dimension in terms of the inclusiveness of the platform model: It tends to attract (under)employed people who want to supplement their principal source of income with supplemental earnings.On the other hand, those workers who use platform work as a first and unique source of income are the most disadvantaged, and are thus less able to derive any benefit from it in terms of flexibility and the fragmentation of work demand.
This accessory and supplementary dimension of earnings is particularly evident in the case of digital microwork platforms.These platforms don't even use the concepts of "work" or "income": On the Crowdville website they encourage people to "earn while having fun…comfortably seated on your sofa at home."On Clickworker, workers "are students and freelancers who generate an additional income with us on a freelance basis."To reinforce its complementary nature in the case of these low-skilled platforms, the gain often takes the form of cashback or vouchers for affiliated stores.It should also be considered that cashback is not convertible into cash and does not even have a fixed value, because it depends on the commercial agreement with the lender companies.Moreover, it is a "credit" system that can only be accessed if certain thresholds are reached, pushing workers to be productive within a given time.
Therefore, the system of rewarding in platforms is more rigid than it seems at first glance, delineating forms of exclusion for those workers who are less readily available and committed, despite the rhetoric of freelancing.Moreover, another serious source of discrimination arises when workers don't have retained any regular employers.For instance, in a platform such as Helpling, clients may cancel their appointment even a few minutes before schedule, incurring risible fees that should be used as the worker's reimbursement for the missed per-formance.However, when the client who reneges their reservation is a well-known contact (in the sense that the involved client and worker had a previous working relationship), some workers may renounce the reimbursement in order to preserve the capital of mutual trust: If [the person] who cancels the appointment is someone with whom I do not have any particular relationship, I accept the monetary reimbursement equal to one hour of work.Otherwise, if I have a continuous relationship with him or her, I say ok, probably it was an accident, it does not matter.(L13, male, 35, Helpling) Similarly, workers manifest the need not to be too "choosy" and selective in their availability when it comes to accepting any type of work.Indeed, due to the "fuzzy" rules regulating the functioning of the platform, they may be afraid that their refusals will be recorded by the system, signaling a low level of motivation that could jeopardize their future chances of employment.The system of "taskification" of rewards typical of many platform jobs seems to increase the servility that connotes the employee-employer relationship, leading workers to adopt strategies and behaviors that often undermine their main goal (i.e., earning money) in favor of a lasting work relationship.This is not dissimilar to what is found in other forms of precarious work, but here it takes on even more worrisome connotations for gig and intermittent work.
Some distinctive aspects emerge concerning highskilled jobs.In this case, the reward system is connoted by the dimension of gamification and challenge among professionals.The underlying competitiveness is sublimated in a "winner-takes-all" logic, which is implemented in numerous platforms such as GoPillar, 99designs, 4CLegal, or Houzz.This competitive dimension is considered a usual condition of these professional markets, starting from the genesis of some platforms, including 99designs, which describes its origins as a creative challenge among young designers.Here, the greatest risk for workers is to make an effort that will not be rewarded, and which therefore can assume the form of unpaid labor.
The existence of such competition has ambiguous effects in terms of inclusiveness: On the one hand, some young professionals underline how the platform has allowed them to enter the market and improve their social capital and networks; on the other hand, the competitive mechanism of the challenge tends always to favor those professionals with greater experience and higher ratings, reproducing existent mechanisms of inequalities according to the "Matthew effect."This approach is also supported by a meritocratic rhetoric and the presumed autonomy of the professionals: Apparently, everyone can make their price in proportion to their own endowment and merits (level of qualification, appreciation for the previous work done).However, on platforms such as Textbrooker, 99design, or Toluna, every performance is also subjected to a quality check before being rewarded.Failure to guarantee the expected standard within a set time involves non-payment, an occurrence that can lead to disputes between the worker and the platform.Indeed, platforms have the option of reviewing times and payments unilaterally based on process analytics: This mechanism is perceived as somewhat unfair and tends to have negative effects in terms of inclusivity and job quality, especially for the most fragile segments of the workforce.
This aspect concurs with resizing the myth of platforms as accessible and open markets for individuals with limited work experience.

Labor Market Security
The dimension of labor market security refers to the extent to which a job is effective in protecting the worker from the risk of unemployment, both in terms of promoting and safeguarding their employability and through the provision of specific protection tools.
The first theme that emerges concerns the formal status of working relationships, since forms of employment on platforms seem to develop within a "grey area."Almost no one has an employment contract, and the terms and conditions of the relationship are often opaque and difficult to understand, especially for poorly educated or foreign low-skilled workers.Somehow, most of the workers interviewed seem to have internalized the ambiguous or insecure nature of an employment relationship via It is a job that I recommend to those with a flexible mentality and to those who are not looking for secu-rity…here there is nothing for sure.(L20, male, 38, Tabbid) In a scenario of strong individualization of risk, platforms deny any responsibility concerning the contractually defined worker-client relationship.Rather, they are oriented to commodify "protection" or "employability," proposing to workers some additional services to reduce their tendency to exit the platform.This is the case of insurance against accidents at work provided by Helpling, or the consultancy service for families that want to regularize their babysitters provided by Le Cicogne.
The possibility of enhancing one's employability on the platform largely depends on the personal branding strategies of each worker (which include caring for their profile, updating information, reputation ratings, feedback, etc.).Successful workers are those who "stand behind [own's] profile a lot" (L20, male, 38, Tabbid), who have "an excellent profile" (L15, female, 54, Le Cicogne), and who recognize that "the platform…works well if you move around…if you interact a lot" (L12, female, 45, Houzz).The unpaid time that workers dedicate to interacting on the platform and taking care of one's "visibility status" becomes an important dimension of inequality in order to ensure job continuity.This may penalize those who work on a platform only part-time and who have only limited time to invest in their personal branding.However, for some fragile individuals, such as women, foreigners, or the elderly, the chance offered by some platforms of omitting personal information can buffer potential prejudices that could otherwise be connected to the workers' personal characteristics: For me, it is an advantage [the fact that on the website there is little personal information about the worker], since if you have to provide a kind of curriculum vitae, your age emerges, and I would be crowded out.(L3,female,57,Helpling) This statement is in line with others that emerged during the interviews, confirming that, compared to traditional job matching mechanisms based on direct personal interactions, the advantage of the platform lies above all in the ease of "opt-in" and of having the possibility of engagement.However, job continuity is not a contractual issue, but is attributed to the subjective worker's capability: A failure is blamed on their own poor self-promotion capacity.To a negative comment from a worker who has not received job offers, Superprof replies: There are many elements that make an advertisement attractive and always at the top of the statistics: a beautiful photo, a verified profile and diploma, recommendations and comments, the possibility of carrying out lessons at the student's home, and above all, in a case like hers, a nice video as evidence of her skills (for example while she is engaged in training).Adding personalization is a key element of distinction, believe me, and sometimes are the small details that make the difference, especially given the huge number of ads that are published every day.
Visibility on the platform becomes a new commodity to be sold to the worker, and some platforms such as Tabbid, Houzz, and Unbuonavvocato sell premium memberships that guarantee greater publicity for their advertisements.The fact that you pay to have more visibility is an important element that invalidates the idea of digital technologies as capable of being more inclusive and meritocratic.Moreover, clients have no way of distinguishing "premium" professionals from "standard" ones, distorting the merit-based principle through which the platform legitimizes its matching capacity.
In the end, this system reproduces logics in which visibility depends on some characteristics of the subjects, like their ability to pay, that do not necessarily reflect their merits and skills.

Quality of the Working Environment
The analysis of the quality of the work environment is divided into job demands, which include the pressure in working times and the risk factors relating to physical health, and job resources, which include autonomy in work, formal and informal learning opportunities, and relationships with colleagues.
About working times and rhythms, it is important to distinguish between project work with lump sum payment and services whose remuneration is based on working time.In the first case, characterizing high-skilled work to a greater extent, the interviewees express an appreciation of the lack of time constraints.As happens in general in freelance work, however, this organizational freedom corresponds to the risk of intensification of work rhythms and the extension of work duties during the hours usually dedicated to activities outside of work.In platform work, this risk is accentuated by the the platform "never closes," and competitiveness is also played out in terms of the speed of response to customer requests.
In the case of low-skilled jobs, compensation is added to this factor based on their working time, detected through the app or platform, resembling forms of "augmented despotism" (Delfanti, 2021).For digital performances, this is frequently associated with digital surveillance modes: For example, the Crowdville platform asks its "crowders" to take screenshots or record the computer screen to demonstrate the activity carried out.The home delivery sector presents some specificities, including the possibility of choosing work shifts, which may also depend on previous performance on the platform.This generates a mechanism that reinforces a "winner-takes-all" logic, not to mention that, in this case, the platform can also exercise its discretion in canceling shifts.
Concerning health and safety risks, there are significant differences between digital and face-to-face work.For the former, the respondents show little awareness of health issues related to exposure to monitors.Instead, there are some concerns regarding the protection of privacy and the risks associated with datafication processes, an issue typical of digital work: There are polls that are done with the webcam, so they detect eye movement to see if you pay attention to the questionnaire.Let's just say I still have a certain reluctance because technology always has limits, especially regarding privacy.(L6, male, 27, Toluna) Vice versa, risks relating to health and safety are prominent among those who carry out face-to-face work.In part, these risks are related to the specificity of the required task and are not dependent on digital intermediation.In these cases, the platforms intervene directly to protect against these risks.This is evident in delivery platforms: The availability of protective devices is one of the criteria adopted by workers to choose the platform through which to operate.Moreover, in the case of in-presence jobs, it is interesting to notice how, since the accountability of the performance is strictly related to the client's evaluation, some workers prefer to act in a "poorer quality" environment so that their effort could be more easily recognized and appreciated: For me [it] is more annoying when I have to work in a clean house, since in this case nothing of my job will be noticed….I prefer a dirty house, it makes me less anxious [about my performance].(L4, female, 22, Helpling) Regarding resources, a central issue concerns autonomy and learning opportunities.In low-skilled work, the activities are fragmented and distributed among people who do not provide specific skills (unbundling of tasks).In highskilled work platforms, this issue is even more complex: The stratification of the external labor market and the de-professionalization process facilitate the positioning of the platforms in the lower range of professional work.This determines a mechanism for self-selection of the most fragile workers: Young people or marginal workers especially may use the platform to gain experience, albeit with low pay.About the possibility of "learning on the job," many platforms introduce some content for self-learning in their blogs, pushing individuals to study and practice useful skills in that sector.The growth mechanisms within the platform are conveyed by the reputational system, which is assumed to allow the recognition of experience and the quality of the skills acquired.
In terms of autonomy, the main criticism concerns the exporting of the human and social capital acquired and building career paths outside the platform, a step which is discouraged by the platforms themselves through lock-in mechanisms.However, it is mainly the pervasive control of the algorithmics that concerns the worker and adds to their greatest discomfort in the work environment: You did not know [which part of the performance] was surveilled…sometimes you were surveilled…then I could not be superficial…at each task, before passing to the following task…a red light is turned on if I performed that task badly….I knew that there was this possibility, but I did not know when and what was [being] surveilled…so it was basically a manager's trick…after three wrong tasks your profile was deactivated for 24 hours, and after three temporary deactivations your profile was cancelled.(L1, male, 28, Clickworker) Another dimension of the quality of the work environment concerns relations with colleagues.In corporate communication, the platforms make extensive use of the concept of "community," borrowed from collaborative economy practices, a rhetoric that generally does not correspond to actual community logic.On the contrary, the interviewees complain about the lack of relationships with other workers and the frequent use of the concept of "alienation" to describe one's own experience is meaningful in this sense.The workers themselves express complaints through online review services such as Trustpilot or Feedaty, since the anonymity guaranteed by the online system of rating may expose workers to unfair evaluations: Online it is very easy to discredit….I've run into several clients who posted negative evaluations on my colleagues who were not punctual in delivering their work or didn't do a good job…it does not take much to be discredited…it is very easy to fail in this "thing."(L20, male, 38, Tabbid)

Conclusions
Our research, adopting a multidimensional approach, contributes to the literature on job quality in platform work.The analysis, while confirming the concerns raised by other studies on the field, highlights the importance of considering the heterogeneity of risks and experiences faced by the different types of platform workers: This view represents a conditio sine qua non to grasp the different nuances of the social inclusivity of platform work.
Our results problematize the thesis of digital plat-forms as a socio-technical system that makes work more inclusive.From our empirical findings, it emerges that the workers who are somehow "socialized" in the labor market, with a protected job and more experience, are those who obtain more benefits from platform work and experience a higher job quality.On the one hand, some patterns of cumulative advantage emerge for selected workers; for others, on the other several adverse employment trajectories across offline and digital work result in an "entrapment" in poor quality jobs.However, different job quality levels and patterns of working conditions are significantly linked to a type of task (online and offline) and the worker's endowment.The type of task and the methods of execution and engagement mediated by the platform greatly affect the quality of the work, differentiating many effects as well as raising possible critical issues and areas of intervention.At the same time, from our analysis, no differences between medium-skilled and low-skilled workers emerge, as already underlined by previous studies.Table 1 summarizes the main findings of this article, highlighting how they affect the relationship between digitalization and social inclusiveness.Among the highly skilled workers who operate mainly in a digital way, many difficulties are related to the challenge-based engagement mechanism that exacerbates competition by lowering the quality of earnings in terms of size and continuity.For the many in professions without a professional order association, the main difficulties are linked to the loss of autonomy in setting the price of one's professional services, while the engagement mechanism based on the challenges strengthens unbalanced and opaque competitive dynamics where the recurring winners marginalize everyone else.These dynamics are reinforced by asymmetrical and unclear logics of distribution and reward on tasks, typical of algorithmic management.For example, the professionals who work with the challenge mechanism must comply with the "winner takes all" principle, which governs the assignment of tasks because the managerial evaluation criteria are not always known.Not to mention the resulting power law that is established, i.e., some workers always work more than others or have better tasks.
Greater visibility on the platform is therefore an essential tool for ensuring employability, and the platforms "sell" tools and spaces to guarantee the personal branding strategies of professionals, especially for insiders and youngest workers who need to develop their reputation and social capital, transforming professional visibility into a commodity that generates further profits for the platform.Moreover, in the case of high-skilled virtual workers, it is like having an office that is "always open," which can result in processes of self-exploitation and a significant increase in workload.
On the other hand, high-skilled professionals who work face-to-face are characterized by a higher individualization of risk because they have to balance the need to stay competitive within the platform, where prices tend to be lower, and the image and the reputation they have to maintain offline to not devalue their competence and professionalism.Some platforms that allow payment to be negotiated and transacted off-platform allow some of these workers to better manage this balance.For them too, visibility is a commodity that is bought on the platform to guarantee more employment opportunities for themselves.However, the interaction and overlap between private and physical workspace, often coinciding with one's home, and online work create greater concerns and risks in terms of privacy or work-life balance.
Low/medium-skilled workers who operate exclusively online are mostly involved in relatively unprofitable micro-tasks that provide "pin money."The ambiguous status of these workers is very problematic in terms of labor security and employability.Moreover, the workers tend to perform highly repetitive and routinized tasks, subject to digital control and surveillance systems.Platform communities, rather than offering a relational dimension, become more like a tool to discipline and control the worker's performance.
Regarding low/medium-skilled workers who offer services on-site, the assessment of the quality of earn-ings is perceived as positive, but only when compared with previous experiences of informal work, very common in these occupations (such as deliveries, cleaning, babysitting, etc.).In this sense, our evidence runs somehow counter to what previously emerged about this category of workers.In such jobs, the need for higher protection against health risks and greater continuity of work is highly perceived.The platform exploits this need by providing protection services, from which it derives an economic profit or an advantaged position over its competitors.Again, the quality of the working environment is heavily determined by the opacity of the algorithmic mechanisms that govern the assignment of workloads and rewards; these algorithms create elements of conflict within the platform, which sometimes give rise to forms of voicing one's concerns and collective organization.
Ultimately, the analysis carried out highlights the risks of a generic and unambiguous regulation of platform work, such as the one currently under discussion in the European Union or tested in some countries, including Italy.Even if common issues emerge, such as access to welfare and the need to reduce the opacity of the algorithmic "black box," a more sectoral and specific approach is needed to make the platform work more inclusive and of higher quality.Such a regulatory approach needs to empower processes of brokering and the representativeness of platform workers.This also means debunking those naive interpretations of platform work that emphasize the myth of disintermediation and more direct interactions between supply and demand.
Our analysis comes not without limitations in terms of the methodological approach and selection criteria adopted.By focusing on Italian workers, this article sheds light on the social inclusiveness of digital work in a socioeconomic context from the Global North: This limits the external validity of the study and, in this sense, a comparative analysis including other countries of the Global South would be welcome.In addition, it could be fruitful to undertake a more in-depth analysis of cooperative platforms, which are characterized by different governance models to those considered in this study.

High-skilled Low/medium skilled Figure
2. Type of platform work and platforms analysed.

Table 1 .
Comparing job quality and inclusivity issues according to the different types of platform work.