Online labor platforms versus temp agencies: What are the differences?

Purpose – The purpose of this paper is to compare online labor platforms (OLPs) such as Upwork, Fiverr, YoungOnes and Temper with traditional temp agencies. At a first glance, OLPs and temp agencies strongly resemble each other while they aim to meet the need for short-term labor of organizations. The authors ask the question how these labor market intermediaries differ on issues such as information technology usage, ways how labor supply and demand are matched and working conditions (e.g.status, payandsocial securityofworkers). Design/methodology/approach – Next to a review of the academic literature, the authors conducted interviews with representatives of six OLPs and temp agencies in the Netherlands as well as a legal specialistinDutchlaborlaw. Findings – The authors found that OLPs and temp agencies differ on several issues. First, although OLPs rely on online marketplaces for matching labor supply and demand, temp agencies generally rely on human matchmakers. Second, although OLPs enable workers and client organizations to initiate transactions themselves, temp agencies employ representatives that do the matching for workers and clients. Third, and as a result, OLPs afford client organizations to almost instantly hire workers on-demand, whereas the flexibility and speed that temp agencies can offer depend on availability and processing capacityofhumanmatchmakers. Originality/value – According to the authors’ knowledge, this paper is the first to compare OLPs and temp agencies and, in doing so, offers academics and practitioners an analytical framework to compare differenttypesoflabormarket intermediaries.


Introduction
How do online labor platforms (OLPs) compare to temp agencies? That is the question we answer in this paper. For decades, organizations are turning to temp agencies for hiring contingent workers. More recently, organizations also started relying on OLPs to meet their needs for short-term labor (Kä ssi and Lehdonvirta, 2018). OLPs refer to organizations that rely on information technology (including mobile applications, artificial intelligence and software algorithms) to matchmake between supply and demand for short-term labor (Kuhn and Maleki, 2017;Meijerink and Keegan, 2019). OLPs come in many different shapes and forms (Duggan et al., 2020): whereas some allow individuals to transact (e.g. Uber that matchmakes between individual consumers and taxi drivers), others matchmake between organizations and individuals (e.g. Twago that supplies businesses with independent consultants). Moreover, although some platform workers perform virtual work in global, online marketplaces (e.g. those of the Upwork, Fiverr and Toptal platforms), others work onsite/offline at the premise of a client (e.g. YoungOnes that supplies temporary workers in domains such as hospitality, logistics or cleaning). For the purpose of this paper, we focus on OLPs that matchmake between supply and demand for contingent labor that is performed on site of a client organization. We do so because these types of OLPs most strongly resemble temp agencies that supply client organizations with a contingent workforce that perform work at the premise of the client organization (Cappelli and Keller, 2013).
At a first glance, OLPs and temp agencies strongly resemble each other. Both: meet the demand for short-term labor; represent a labor market intermediary that enables client organizations to hire contingent workers; and charge a fee for their intermediation services (Duggan et al., 2020;Meijerink and Keegan, 2019).
This builds the question whether and how OLPs actually differ from temp agencies. To our knowledge, empirical studies that answer this question are nonexistent. This is surprising because such studies would offer hiring organizations, and their (human resource) managers, an overview for comparing labor market intermediaries. To make up for this, we conducted interviews with representatives of six OLPs and temp agencies in the Netherlands as well as a legal specialist in labor law to: uncover how both types of labor market intermediaries differ; and derive a comparative overview [1]. This is not to say that we advocate for the use of contingent labor by hiring organizations. In fact, we argue that our overview is useful for hiring organizations to decide whether they want to hire contingent labor in the first place, and only if so, whether to want to rely on temp agencies or OLPs.

How online labor platforms differ from temp agencies
The interviews conducted reveal that OLPs and temp agencies differ in a number of ways (Table 1). First, they differ in terms of how demand and supply for labor are matched. Although OLPs almost exclusively rely on information and communication technology (ICT) to facilitate matchmaking, temp agencies mostly work with what we call "human matchmakers" (e.g. employee representatives of the temp agency) who do the actual intermediation between supply and demand for short-term labor. This is not to say that temp agencies refrain from using information technologies. In fact, temp agencies use ICT to collect and store information about their contingent workforce (e.g. about qualifications, work experience and current assignments) that human matchmakers rely on to supply contingent workers to a client organization. OLPs differ here while they do not employ human matchmakers and, instead, operate an online marketplace where labor supply and demand meet.
Second, the latter implies that workers and hiring organizationsin the case of OLPsare given more responsibility (and freedom) to initiate transactions themselves. This works as follows: by means of a Web browser or smartphone application, hiring organizations can post a call for a short-term assignment that individual workers can subscribe to. In turn, the hiring organization can choose itself which of the workers, who subscribed to the call online, it wants to hire. Some of the OLP representatives that we interviewed indicated that their online platform allows hiring organizations to create a pool of "favorite" workers who have priority or exclusive access to calls posted by a hiring organization. Temp agencies differ here, as hiring organizations need to turn to a human matchmaker employed by the temp agency to request for a contingent worker. When relying on a temp agency, hiring organizations may not decide themselves which contingent worker to hire. Instead, this is done by the human matchmaker who decides which individual worker to supply to a hiring organization.
Third, although temp agencies and OLPs both offer labor flexibility to hiring organizations, they differ in the degree of flexibility that is provided. This follows from differences in the matchmaking approaches of OLPs and temp agencies, which we described earlier. As OLPs facilitate matchmaking via online marketplaces, and thereby devolve hiring decisions to workers and client organizations, they can be more flexible than temp agencies. For instance, although temp agencies may be closed during the weekends and evenings (when human matchmakers enjoy leisure time), the online marketplaces of OLPs are accessible to workers and hiring organizations on a 24/7 basis. Moreover, OLPs allow their clients to reach out to a worker directly, thereby enabling them to hire contingent workers instantly and on-demand (depending on how quick a suitable candidate responds to a hiring organization's call, of course). As such, although temp agencies offer flexibility, OLPsat least in theoryoffer what we call hyperflexiblity (i.e. close to instant, on-demand hiring of a contingent workers without the interference of a human matchmaker). The result of this all is that OLPsin comparison to temp agenciesare more likely to limit information asymmetries vis-à-vis client organizations (as the latter has a more complete overview of potential job candidates) and reduce operating costs as OLPs do not need to hire (many) human matchmakers that temp agencies do work with. Accordingly, temp agencies are more likely to supply workers that perform long-term work assignments (e.g. that last weeks or months) as these bring enough revenue that justify hiring humans who matchmake between supply and demand for contingent labor. On the contrary, we found that OLPs work "day laborers" (Cappelli and Keller, 2013) who perform short-term "gigs" that last no longer than a day. As noted by one of the platform representatives interviewed, these shortterm "gigs" tend to be 6.5 h in length, on average.
Taken together, the above implies that OLPs empower workers and clients to initiate, handle and control transactions themselves. This is why OLPs are especially suited for catering to the demand of hiring organizations for hyperflexible, contingent labor. Essentially, OLPs devolve responsibilities (that the human matchmakers of temp agencies perform) to workers and client organizations. As such, OLPs offer "low frills services" to its users, which lowers operating costs to the platform firm but increases administrative burdens for client organizations (that need to, e.g. draft and publish online calls, select workers and manage a pool of favorite workers themselves). Accordingly, client organizations that wish to be supported by, and outsource administrative tasks to, a labor market intermediary can best turn to temp agencies.

Differences across online labor platforms
Although OLPs differ from temp agencies, our research revealed that OLPs themselves also differ in terms of their relationship with contingent workers. Here, we found two types of OLPs: those that institute an employment relationship with their workers (hereafter called: co-employment platforms) and those that work independent contractors (hereafter called: freelance platforms). Co-employment and freelance platforms differ on a number of dimensions (Table 1). First, they differ in terms of pay (flexibility) offered to workers. As noted by the interviewee representatives of the co-employment platforms, their workers are employed by the platform and, therefore, subject to collective labor agreements of the platform firm and/or that of a client organization. Accordingly, co-employment platforms offer their contingent workforce a more or less fixed hourly pay rate (to which clients may add a surcharge in times of labor scarcity). Freelance platforms instead work with solo selfemployed workers who are nominally one-manned business and therefore free to set their hourly rate. Depending on how the matchmaking process is technically organized, it is either the freelance worker or client organization that makes a proposal on the pay level. Therefore, and depending on the bargaining power of individual workers, we expect varying levels of pay rates on freelance platforms. In fact, some of the Web applications of the freelance platforms we studied allow independent contractors and hiring organizations to negotiate pay rates by making counterbids. At the same time, to avoid "races to the bottom," some of the freelance platforms' digital interfaces do not allow workers and hiring organizations to set hourly pay rates below a minimum level.
Second, OLPs that employ their workforce are obligedby (Dutch) lawto safeguard workers' occupational safety. Among others, this means co-employment platforms are obliged to assess whether working conditions at the client organization's premise are safe. In addition, workers of co-employment platforms are entitled to disability insurance. Such legal obligations do not exist for freelance platforms. In fact, freelance workers have to organize a range of social security schemes (e.g. paid sick leave, paid parental leave or disability insurance) themselves. These can be purchased from for-profit insurance companies or be organized collectively by freelancers (e.g. by establishing a "Bread fund"). In addition, freelance workers have to make up for employment security themselves as well. This is not to say however that workers of co-employment platforms are not necessarily better off in comparison to freelance workers. For instance, throughout our interviews, we found that workers of Dutch co-employment platformsin the first 78 weeks of employmentcan be offered an infinite number of short-term contracts (i.e. daily contracts) by the platform firm without a guarantee of being rehired, not earning an income in case they fall ill (NB: the same applies when the client organization cancels the assignment four days prior to its start) and losing their job (i.e. contract with co-employment platform) if the demand for labor, i.e. the work assignment, at the hiring organizations ends altogether.
Third, co-employment and freelance platforms differ in terms of replacing workers who do not show up for work (i.e. so called "no-shows"). In the case of co-employment platforms, the responsibility for replacing such workers lays with the platform firm. This responsibility follows from the contracts that co-employment platforms engage in with hiring organizations, specifying that the OLP is responsible for supplying their clients with an appropriate (and qualified) number of workers. Freelance platforms do not have this obligation, which instead is devolved to freelancers who are for finding replacement themselves. Some freelance platforms that we studied are fining workers for no-shows and/ or enable workers to have themselves replaced. They do so by building features into their online platforms that allow workers to subcontract work to other freelancers and allow hiring organizations to evaluate the performance of the replacement worker. According to one of the platform representatives we interviewed, about 75% of the freelancers who performed more than 50 "gigs" via the platform engaged in subcontracting at least once.
Finally, OLPs differ in terms of integration with another (incumbent) labor market intermediary. Throughout our interviews, we found that freelance platforms tend to be stand-alone organizations with a sole purpose to coordinate an online marketplace where supply and demand for contingent labor meet. Although some co-employment platforms are similar to freelance platforms in that they are stand-alone organizations others are (horizontally) integrated with a traditional temp agency. Temp agencies establish OLPs to use such digital platforms as a software tool to optimize the matchmaking process performed by human matchmakers and/or to extend the portfolio of intermediation services for meeting a more diverse set of client needs. Whether OLPs are standalone or combined with the services of a temp agency, these have implications for the career prospects of individual workers. In the case of the latter, workers can more easily access a broader set of work assignments and have a more realistic outlook for a stable contract (and at times, a permanent job).

Conclusion
As labor market intermediaries that matchmake between supply and demand for contingent labor, OLPs and temp agencies tend to resemble one another. At the same time, they differ while OLPsin comparison to temp agenciesenable a hyperflex supply of contingent workers, involve little control/oversight by human matchmakers, create less information asymmetries vis-à-vis client organizations, offer more freedom and empowerment to workers and clients (meaning they cannot outsource administrative tasks to the intermediary) and lower operating costs (to the labor market intermediary). As a collective, OLPs can be further differentiated in those that engage with workers who are independent contractors (i.e. freelance platforms) versus those who are employees to the platform firm (i.e. co-employment platforms), which has implications for the pay, social security and responsibilities of workers.
In our study, we primarily examined differences at the level of the organization. This leaves many questions unanswered, especially those that concern the interests, attitudes and behaviors of individual workers. For instance, although freelance platform representatives indicated their workforce to be more entrepreneurial (given their self-employed status), we do not know whether this is actually the case. As such, future research may want to answer questions such as: are freelance workers more pro-active than co-employed workers? To what extent do workers of freelance platforms subcontract platform work to others, and what type of work assignments do they tend to subcontract? Are freelance workers less likely to engage in "no-shows" than their co-employed counterparts? To what extent do freelance and co-employed workers differ in the level of (experienced) job security, pension savings and insurances? To which extent do freelance workers experience whether freelance platforms ensure safe working conditions at client organizations, or do workers make up for this themselves?
Notwithstanding the need for more research into employee-level differences, we have given an overview of the similarities and differences between temp agencies and OLPs. In doing so, we hope that it supports managers human resources in making a decision if, and under what conditions, they wish to rely on temp agencies and/or OLPs for hiring contingent workers.

Note
1. For the purpose of this article, we do not delve into the complexity associated with the legal status of platform workers. Instead, we compare OLPs and temp agencies in terms of service provision, business processes, human resource management activities and relationships with workers and client organizations.