Key workforce virtualization challenges
In previous posts (here and here), we put forth a vision of an enterprise workforce model that is primarily driven through market forces rather than through traditional static organizational hierarchies. We further outlined our ongoing research projects related to the umbrella of a workforce virtualization platform.
The basic notion of workforce virtualization and crowdsourcing isn’t inherently difficult to grasp. However, making such a model scale and work in an enterprise setting can present several issues. In this post, we describe the key challenges that are relevant to a workforce virtualization platform, followed by our plan to explore and address these problems in our work at the Labs.
Security and trust: Data security has become a major issue for companies in the digital era. To protect valuable and sensitive data, enterprises have instituted strict security and compliance policies, although it is not uncommon to find these policies differ among departments within a large organization. How can external crowds be trusted to work on- demand for an enterprise, without adding too much overhead to work processes?
In addition to the standard taxonomy of workforce as being internal or external to an enterprise, a third category may be necessary, one that sits at the intersection between internal and external crowd. We call this category curated crowd, and it is comprised of workers who are not technically employed by a company, but are vetted fully for security compliance and who can be trusted by the company to carry out its tasks.
Members of this curated crowd can include contractor channels, company alums and vetted workers from external service providers. Companies will likely to need to expand or augment compliance and security policies for this group. Similarly, jobs will need to be strictly classified in accordance with the compliance model before being distributed to this group.
This is where workforce virtualization can be of significant value by:
Creating validated and standardized screening processes that help to vet crowd members and ensure they are on-boarded with full compliance.
Providing a standard classification scheme that automatically tags jobs with the required security clearance.
Ensuring that the jobs are targeted only within a subset of the crowd members that have the appropriate access.
Reliability: A hallmark of workforce virtualization is its ability to scale with respect to workload demands. This flexibility does come with a risk. When a task is posted to the crowd, what level of confidence can a poster have that a crowd member with the right skills will step up to apply? How can the poster set reasonable expectations with regard to the quality and timeliness of the task to be completed? These reliability issues will become increasingly important as the on-demand, market-driven work allocation model becomes pervasive.
For our workforce virtualization platform, we are using machine learning techniques to create predictive models that identify factors that are strongly correlated with task completion and quality. These factors are related both to the way jobs are described and specified, as well as the available talents and expertise of the members of the crowd.
Such models will not only help set the appropriate expectations for the job poster, but will also help optimize the crowd targeting and notification schemes to maximize the chances of work up-take and high quality completion. In essence, we hope to transform today’s predominant ‘post and hope’ model of crowdsourcing to a more deterministic ‘post and expect’ model. For more information, see our related blog post.
Incentive and motivation: A complementary issue to the reliability challenge is how to create proper, sustained incentives to motivate crowd members to continue to produce work on-time and with a high level of quality. We have found reputation to be a key incentive mechanism for crowdsourcing. One of our studies showed that reputation can often be a stronger incentive to online workers than monetary compensation, as working for a reputable employer and receiving good feedback can maximize their future hiring prospects.
Since reputation plays a direct role in vendor labor markets (e.g., UpWorks), vendor platforms are typically designed to provide an explicit means to capture and present reputations. However, this is not the case for an internal enterprise workforce, where vitae may only contain self-reported skills and work feedback may only be shared between an employee and manager. Accurately capturing people’s skill sets and validating them on an ongoing basis will be important to drive proper incentives for internal crowds.
Coordination and context transfer: The notion of teams and team membership will become much more fluid in the context of a virtual labor market where workers are primarily engaged on-demand. While this flexibility offers benefits for both the company (less overhead) and the workers (pick and choose projects based on interest), it does raise the following issues:
How can context be transferred to and from transient team members to operate as teams that are rapidly formed and re-formed in both small and large projects?
How can awareness and cohesion be maintained among a widely distributed team of crowd workers who do not generally know each other well?
Here, our workforce virtualization platform builds on and extends our existing research by bringing together collaboration, enterprise social and mobile technologies to create “intelligent digital collaboration work processes.” These processes streamline context transfer, coordinate work hand-off and improve communication among members of a transient team.
In the posts that follow, we will discuss projects under our workforce virtualization platform, starting with talent brokering.
For more information about crowdsourced workforce virtualization, please mail Alex Kass at email@example.com. We would like to thank David Q. Sun for his contribution to this post.