Cloud to crowd talk to COST Virtual work Bucharest 2014
Post on 19-May-2015
DESCRIPTIONThe market place for work and work opportunities is changing, mediated by online markets and exchange. Here I present findings from an EC study on the trends, challenges and opportunities coming from crowdsourcing (crowd flower, AMT, Samasource etc), cloud work forces, and global freelance exchanges (ODesk, Elance etc). The presentation explores how these services have been developed, shaped, and used, and issues that policy makers should be aware of.
- 1. From Crowd to Cloud: The Present and Future of Work in the Network Economy the role of Online Work Exchanges James Stewart University of Edinburgh email@example.com Virtual Work, Bucharest March 2014 This work was funded by and conducted at the JRC-IPTS, European Commission ICT4EMPL "The Future of Work" This presentation does not represent the views of the European Commission
2. My concern from policy? 207/09/14 3. Online work exchanges: New ways of finding work, finding workers, being employed and getting work done Not: Telework, free labour etc 307/09/14 4. Potential Policy Opportunties Employment-related policy (such as labour market reform, temporary work, youth employment and training, entrepreneurship, self- employment, flexible working, access to work), Balance of industry and welfare (e.g.flexicurity) Skills policy Digital skills and access, Mobility programmes Enterprise and Markets SMEs and microenterprises, microfinance Regional Development, Financial regulation, The Single Market in services and employment. Job Creation Social Policy Social inclusion programmes Social cohesion programmes Public service delivery International development 5. 10+ years on Bates and Huws 2002 'eworker' estimates for 2000 in Europe (EMERGENCE Project) 9 million eWorkers 3.7m multi-locational eWorkers 810,000 teleworkers 1.45m eLancers 3m+ eEnabled self employed Most not working through online exchanges 507/09/14 6. STS Agendas Software Infrastructures Scale > Big Data Humans and computers ever more tightly entangled Algorithmic matching Google search, adwords, social media (Facebook etc) Social computing and classification ICT and Work Replacement of People by IT End of Work Telework CSCW Mobile Work Open Source and voluntary labour Call centres Globalisation, BPO and off-shoring Fragmentation of labour solidarity Hollowing out of white collar work Crowdsourcing for free labour Virtual Work07/09/14 7. 707/09/14 Clickworker Elancer Turker Cloudworker 8. 807/09/14 9. 907/09/14 10. 1007/09/14 11. 1107/09/14 12. 1207/09/14 13. 1307/09/14 14. 1407/09/14 15. 1507/09/14 16. 1607/09/14 17. 1707/09/14 18. 1807/09/14 19. 1907/09/14 20. On-site Location of Work 2007/09/14 Off-site 21. Who works through these systems? Working and Unemployed Students gaining experience, reputation or spare cash People with families looking for flexibility Disabled housebound Rural dwellers Middle-aged restarters Retired people supplementing pension People in traditional Freelance occupations Designers Translators Accountants Programmers Small Business Hipsters Professionals in South Asia and other emerging economies Microbusiness and freelancers buying services from others 2107/09/14 22. Types of tasks (after Frei) Size, org,pay example Microtasks High volume; low pay per task; automated Transcription, classifying, price search, find simple info 'Macro'-tasks High volume, low pay, automated Product review, simple testing, simple info collecting (e.g marketing) Simple projects Low volume, single tasks, with skill and moderate pay. Direct contact Design a website Do accounts Write a presentation Design a logo Complex projects Single projects, high pay, often multiple people, direct contact Scientific challenges Algorithm design Complex research 2207/09/14 23. TWO GENERIC OPERATING AND BUSINESS MODELS 2307/09/14 24. Crowd Crowdsourced Microwork model 2407/09/14 Crowd Platforms Task Managers and Resellers Local intermediaries Curating the crowd Clients (direct integration) Clients (ad hoc) APIs Crowd self- organising Microwork design Workflow integration Quality Matching Recruitment Worker interface (motivation, quality, payment), APIs, BPO workforce 25. Freelancer Marketplace model Trust Efficiency Transparency 07/09/14 ContractorsClients Intermediaries and market makers Large clients Matching Payments Quality Dispute management Support Work platforms Value added services Competitions Contracts Rent-a-crowds Teams Access to clients Support and training Access to Cntractors Value added services Tax Training Mentors Resources 26. Existing research concerns and results 2607/09/14 27. Research on online exchanges Crowdsourcing business models +some critical user studies (Brabham 2010, 2011, 2013). Elancing from an HR perspective (Aguinis and Lawal 2013) Microtask platform use e.g. in scientific experiments (Iperitos 2008,210a, 2010b) Labour economics perspective (Agrawal et al 2013) Virtual labour Huws 2003; Scholz 2012; Kleemann and Vo, 2008; Huws 2013; Holts (2013) Caraway (2010) ) Legal issues (Felstiner (2011) Microworker identity (Lehdonvirta and Mezier (2013) Microworker empowerment - Turkopticon (Irani and Silberman 2013). 2707/09/14 28. Open the black box of job search (Petrongolo and Pissarides, 2001; Marchal et al 2007). formal and informal information channels (e.g. Granovetter 1974) role of intermediaries whose work is to match vacancies sellers and buyers (see Marchal et al 2007). 2807/09/14 29. Literature Practitioner Interviews Short Cases Inductive In depth qualitative cases and analysis Outsourced to Warwick university 2907/09/14 ICT4EMPL 30. 3007/09/14 31. 3107/09/14 Specialist Generalist Global Proz (Translators) Microtask Elance; AMT National/ Regional Trada (optimizers) PPH; Clickworker Local Rated People (domestic trades) Slivers (Social care) (Local listings) 32. 3207/09/14 12 000 000 $16 000 000 000 33. Online freelance sites: 12m worldwide (World Bank estimates from adding top 3 elancer sites, neglects multiple membership) Elance 2.3+ million registered users 715k in US, 359k India, 80k UK $200m elancer earnings. 48% say main source of income Odesk Matched 35m hours of work in 2012 workers in 179 countries $360m earned. 2/3 workers >50% of family income Freelancer claims 7m registered workers, 4.5m completed projects Staffing industry Analysts estimate $1bn value in 2012 ($2bn 2014) Proz 600 000 registered translators, 20 000 paying members Trada 10000s of users 300 regular workers 3307/09/14 34. Microwork Numbers Clickworker 300 000 Clickworkers 1/3 Germany, 1/3 rest of Europe 1/3 North America Crowdflower Claim a crowd of over 2 million 4m human judgments per day 959,582,877 judgements (8/6/2013) Amazon Mechanical Turk The only numbers that we share regarding our Worker population are these two: Over 500K registered Workers from over 190 countries worldwide. Jan 2011 Jobs 1cent-$10 Iperitos, using 2008 data Turkers are younger. Turkers are mainly female. Turkers have lower income. of the general 3407/09/14 35. Rates of Pay Clickworker 8-9euros/hour Elance minimum $3/hour Mturk $0.10 1-2min HIT Penny HITS - for the desperate, adjusted to local (low wage) labour rates. X Time worked Clickworker most people earn less than $300/month Trada top earners on >$5 000 month full time Odesk 2/3 earn over 50% of family income. Proz full time professional occupation 3507/09/14 36. In Europe There are microworkers (culturally specific microtasks) There are online freelancers etc How many? Millions How could we count them? 3607/09/14 37. VALUE AND RISKS 3707/09/14 38. Value and risks for clients New Value Only solution On demand Speed Scalable Exploit crowd effects Analytics Assured service High service quality for specific work HR Lower HR search costs No/low employment costs or obligations Greater selection of workers Access to global pool of talent+ global wage rates Risks Low control Too much choice Lower quality Disadvantages of non- permanent staff Job specification Privacy and confidentiality Complexity of some microwork 38 07/09/14 39. Suggested value and risk for workers Free, Cheap, 'exploitation 'insecurity' Flexible, 'freedom', 'opportunity' Non-economic Flexibility Self employment Work-life balance Life course Try out, and learn new skills Something to do 07/09/14 Economic Makes independent work more feasible Re-enter labour market Extra cash Supplement main income Access to (global) clients Transparency of markets - Trust in market Build a portfolio of clients. Specialisation (Malone et al) Access to work for excluded Tools for productivity Develop skills and employability 40. Key business innovations How did these exchanges become this way and where are they going? 4007/09/14 41. Intermediaries need to attract and keep customers in a sceptical and competitive market Multiple Quality Systems Market management 4107/09/14 Trust and Quality 42. E-REPUTATION, RATINGS AND QUALIFICATIONS 4207/09/14 43. 4307/09/14 44. E- Reputation in the marketplace Workers Rating on every job Algorithmic Reputation Calculation Computer generated story 44 07/09/14 Failure can be terminal (compare with Tripadvisor) Dispute resolution 45. Reputation the Client/Buyers 4507/09/14 A Rating by supplier Bluebook comments B No rating Off-platform fora or hacktivism (Turkopticon) Depends on the business model of intermediary, and the balance of the market. 46. Qualifications and Validated workers Real-life Qualifications Virtual Qualification 'Gold' tests Access to good work (promotion) 4607/09/14 47. 4707/09/14 48. Disputes over qualification tests I was up until 1.00am completing the Qualify Author section. This morning I checked and they only scored me 58%. I felt this was a good article which fitted the brief and used SEO [Search engine optimization] words so I have emailed them. I'm pretty sure it was because the computer timed me (even though I didn't see anywhere which said this was a timed section) out as I'm 4 stars on text broker and got 100% on the 72 questions. Had an email back. They agreed the article was better than the 58% and have changed my mark. He explained it was a little off topic and doesn't need to have witty remarks and more about keywords. At least I can work for them now. (moneysavingexpert.com, 2012) 4807/09/14 49. Transferability 4907/09/14 Private marketplaces privatised qualification & reputation scores Non-Transferable Key to business model keep workers on the platform BUT User-driven portability Posting reputation scores outside the site (Linkedin) Growing recognition. 50. THE CROWD: INDIVIDUALIST OR COLLECTIVIST? 5007/09/14 51. Worker Support and networking 5107/09/14 A community from a crowd 52. 5207/09/14 53. Emerging e-working ecosystem Ecosystem of competitive services and user-generated support User Organisation activities Worker support from the exchange Worker mutual support in third spaces (e.g. http://cloudmebaby.com/mturkblog/ moneysavingexpert.com ) Worker-led creation of closed shops Interventions to add support tools Platforms as tools for building teams, subcontracting and building businesses 53 54. Crowds become communities that: Support techno-socialisation into the practices and logic of work though online exchanges how to perform in this new eco-system. Subversion of exchange logic 5407/09/14 55. Policy Questions and Challenges 5507/09/14 56. Enterprise - SMEs How much start-up, SME and microbusiness growth will/could this type of work exchange create? In what sectors? Countries? How many people might this growth provide work for? For how much of their time? For what overall income? In which sectors of the workforce What kind of support could be given to promote access to services through these exchanges? Awareness; certification; Legal consistency across Europe? Identity proof? 56 57. Employment and Employability Does working through exchanges and crowdsourcing platforms provide: -Access to work -Income -Entry to work -Opportunity to build skills For whom, in what sectors, in what countries? Who does it exclude? Which sort of people are better prepared for this work, and why? How can we support people to develop freelance skills and work ? Could this be integrated with public employment services, how, where? 57 58. Export of Jobs? 5807/09/14 59. Welfare issues Self-employed: low earnings, discontinuous work, low skills, long and non-standard working hours, the high incidence of industrial accidents and work-related health problems Do we want to encourage this? Or do we just have to cope with this reality? Policy concerns -Policy barriers to autonomous workers -The EES not adapting to atypical work 59 07/09/14 60. Platform Support? Why do these platforms not exist in some sectors/countries/regions? Are what rate are they growing? What would be the impact on employment (growth, reduction) and wages if they did exist? If they are desirable to support growth and/or employment, What can public policy makers and public services do to: Stimulate growth in these platforms focused in sectors, regions that will most benefit/most in need? Integration of e-government services with platforms. What are the Risks of promoting this type of working? Insecure work, exploitation, resistance from social partners 6007/09/14 61. 6107/09/14 62. Potential policy interest Reactive Training and skills Constraints on autonomous workers in Europe Protection of workers Flexibility in hiring freelancers Adapt e-gov systems Long term Welfare and Economic issues Proactive 1.Model and ideas for social innovation: Public services PES Public service delivery 2. Active promotion of work these working patterns: opportunities for New jobs Entrepreneurs SMEs Digital economy Excluded citizens Employment transitions 6207/09/14 63. Discuss 6307/09/14 64. Summary Explore: Innovation and investment necessary to build confidence in platforms, and attract clients and contractors Conditions of work, including the pressures of working for and within software machines Potential socio-economic impact 6407/09/14 65. More info on the JRC-IPTS research JRC-IPTS Employability-The Future of Work 6507/09/14 Or James Stewart firstname.lastname@example.org