The future of learning and skills matching from the lens of online platform markets.
The rise of online platform work
A growing number of people are earning some or all of their income from work mediated through digital platforms. In what is often referred to as “crowdwork” or ‘gig work’ or ‘online freelance work’, self-employed workers increasingly work remotely for clients discovered via online labour platforms, in projects ranging from repetitive data entry tasks to specialised software development and creative work. Emerging research suggests that such ‘crowdwork’ is an increasingly important new non-standard form of employment around the world. Platform-mediated work is now the main source of income for as many as 2% of adults in 14 EU Member States.
Although crowdwork represents only a small part of the overall labour market, it still represents the most visible example of technology shifts taking place in the wider labour market, such as growing self-employment and contingent work, customer feedback in lieu of line management, teleworking and virtual teamwork and firms’ use of data analytics and algorithmic management. The new policy challenges posed by crowdwork thus have potentially broader implications for European skills policy.
How do crowdworkers develop and match their skills to gigs?
The Cedefop CrowdLearn research project, carried out in 2017-19, is the first in-depth study to have examined how EU crowdworkers develop and match their skills with gig work, a view to drawing lessons for European skills and education policy. The CrowdLearn project has revealed some striking findings that offer a glimpse into the future of learning and work.
Crowdworkers’ skills development
A novel skills typology, developed as part of the project (Figure 1), reveals that crowdworkers most commonly develop their technical/core skills and specifically their digital skills. However, a unique blend of ‘entrepreneurial’, ‘self-branding’, ‘communication’ and ‘organisational’ skills as well as focused personal dispositions are found to be a particularly important prerequisite for being successful in online platform markets.
Figure 1: Typology of skills developed in crowdwork
Source: Cedefop’s CrowdLearn project; based on a full typology of 123 distinct skills learnt before joining platform work and 89 distinct skills learned during crowdwork.
On-the-job skill development (on a weekly or even daily basis) is found to be a key feature of crowdwork. Platform workers are more likely to develop their transversal skills, for instance how to communicate with clients, ways to self-organise or their professional attributes, as well as platform- and freelancing-specific skills (Figure 2). At the same time, they are keen to take active steps to mitigate skill gaps in their technical and communication skills (Figure 3). In general, crowdworkers tend to rely on just-in-time or bite-sized acquisition of knowledge and skills required to complete specific work tasks. (Free) online video tutorials and websites are among the most important resources of informal learning. Formal training courses offered by learning providers are, by contrast, perceived as too time-consuming and generic.
Platform companies have also been actively developing initiatives for skills development, including fostering commercial partnerships and joint courses with training providers and facilitating peer-to-peer courses and digital social spaces for learning (e.g. community forums). However, Cedefop’s CrowdLearn study reveals that platform companies are generally wary of getting directly involved in the field of skills development, as it could risk them being legally classified as employers; this fosters a potential underinvestment in crowdworkers’ skill formation.
Figure 2: Skills crowdworkers develop through their work on online platforms
Source: Cedefop’s CrowdLearn dataset
Figure 3: Crowdworkers' learning focus / skill gaps addressed in the past month
Skill use in crowdwork
A key additional finding of the study is that crowdworkers often feel that they underuse their full set of skills in their work. This sometimes happens as they accept to take a lower-skilled task due to financial needs or as an entry step for building their job portfolio. But others also purposefully select a less challenging job due to interest for a project and wish to develop a new skill area. Inefficient matching also takes place due to platform clients often posting poor quality job vacancies.
Skills matching in crowdwork
Skills matching in online platform work takes place in a number of ways, such as platforms
- publishing lists of their most-in-demand skills by their clients;
- employing matching algorithms in order to endorse workers for different projects;
- vetting new freelancers prior to their entry;
- promoting client feedback/reputation scores as a signal of crowdworker trustworthiness.
Qualifications obtained through formal education, the usual means of signaling and screening workers in labour markets, are instead considered a weaker signal of skills in crowdwork. Platform-administered digital skill tests and certifications/badges are also available, and for around 2/3 out of 10 crowdworkers they were helpful in getting additional gigs. However, their value in helping crowdworkers enter and succeed in the market was perceived less positively by 6/7 out of 10 such workers (Figure 4), and evidence from the qualitative interviews with platform owners and workers carried out as part of the project corroborates that many clients prefer crowdworker’s job profiles and feedback scores over such tests, which generally fail to signal workers’ overall credibility and can be easily manipulated or become outdated.
Figure 4: Crowdworkers' attitudes towards and utilisation of in-platform skills tests
Source: Cedefop’s CrowdLearn dataset
A new policy framework for platform work?
Promoting integration and mobility
Success in online platform work is mostly dependent on workers’ track record of trustworthiness, which may erect barriers to the integration of newcomers in online platform markets. Platform companies and policy makers collaborating on developing a programme of "micro-internships" is a novel idea espoused by the Crowdlearn study. In such a programme platforms’ clients will be offered a subsidised rate on crowdworkers who lack previous platform work experience. In exchange, the clients would be required to provide the workers with private feedback for skill development and public feedback to validate the workers’ skills and trustworthiness.
In general, policymakers should work more closely with platforms and their clients so as to induce them to offer developmental as opposed to evaluative feedback on crowdworkers’ skills proficiency, the latter often used for punitive purposes. Another issue of concern for policy is the limited inter-platform portability of skills and client feedback posted on crowdworkers’ profiles, which hinders their mobility. But policy interventions to facilitate the transfer of ratings are likely to stumble upon platform-specificities and lack of interoperability in technical architectures, data protection constraints and non-commensurate skills taxonomies, which render standardisation difficult.
Improving education and training
In order to meet crowdworkers’ “just-in-time” learning needs, adult vocational learning providers will also have to consider investing in the development of short and focused online courses, tutorials and workshops. Trade unions could also play a key role in this regard, drawing on their resources, expertise, and partnerships, to provide training opportunities to freelancers, particularly on issues of social security, taxation, and self-employed workers’ rights.
But ultimately initial vocational and higher education providers should focus on developing peoples’ digital and entrepreneurial skills. Despite the low value attached to formal qualifications and skill certificates, crowdworkers do make significant use of skills developed prior to their entry to the online platform market, computer literacy in particular, highlighting the critical importance of investment in high quality digital upskilling training.
An enhanced focus on developing self-regulatory learning skills and mindsets among learners is also critical, such as the ability to be strategic and dynamic in identifying one’s own learning goals, and being proactive in seeking feedback. These are a fundamental skillset to have in the 21st century, best developed before entering employment, whether it is online or offline.
 Pesole, A., Urzi Brancati, C., Fernández Macías, E., Biagi, F., & González Vázquez, I. (2018). Platform workers in Europe evidence from the COLLEEM survey. Luxembourg: Publications Office of the European Commission
 The CrowdLearn project is a cornerstone of Cedefop’s Digitalisation, AI and future of work thematic activity. It was carried out by a research team comprised of the University of Oxford Internet Institute (led by Prof. Vili Lehdonvirta) and the Copenhagen Business School (Prof. Anoush Margaryan). Evidence was collected from in-depth interviews with platform workers, platform owners and policy stakeholders and a new dataset of 1000 crowdworkers from four popular online labour platforms. All reports and deliverables are available for download on the project website.
 As also confirmed by Margaryan, A. (2019a). Workplace learning in crowdwork. Journal of Workplace Learning, 31(4), p. 250-273.
 As testament to this, Upwork removed most of its skill tests from its platform in 2019, only retaining a ‘readiness test’ for newcomers. Other platforms however take success in skills courses offered by them into account when ranking its workers into different levels of proficiency.