Experimenting with the Workforce Virtualization toolkit
In Part 1 of this series, we introduced our perspective on crowdsourced workforce virtualization as an alternative workforce model that promises to create a nimbler enterprise. A transformation as big as this—from a traditional workforce structure to a much heavier reliance on crowdsourcing and dynamic labor markets—can be challenging to carry out. It requires the mastery of new workforce mindsets and methodologies, along with the adoption of technologies from a rapidly evolving landscape. Various forms of work also require distinct tools and techniques.
Most enterprises will begin slowly, identifying simple tasks to virtualize before moving on to more challenging forms of work to assign to the crowd as their mastery of workforce virtualization grows. At the Labs, we are exploring what first steps are possible and next steps look fruitful to realize the full potential of this transformation. (See Figure 1.)
|First steps||What is possible now with existing technologies?||
|Next steps||What could be feasible soon?||
Figure 1: Steps for transitioning to workforce virtualization
To answer these questions, we have been studying the landscape of relevant commercial technologies, running some hands-on experiments, and building out tools and techniques for addressing the issues uncovered by those experiments.
Evaluating the technology landscape
The crowdsourcing space has become broad, which can make it confusing to track and analyze the dozens of commercial vendors and conceptual approaches. To help evaluate the options and formulate specific strategies for enterprises, we have developed a four-dimensional framework for understanding vendors and approaches. (See Figure 2.)
While the basic principles of crowdsourcing and workforce virtualization are not inherently complex, the framework shows that there are many interesting variations in how these principles can be combined to support different kinds of work. This is one reason why there has been a proliferation of commercial platforms, as shown in Figure 3.
|For the enterprise|
|Micro-task labor markets||Public digital talent markets that allow task owners to stream many very small tasks (such as classifying a photograph or sentence). Support human-in-the-loop computation.||
|Skilled freelancer labor markets||Public digital talent markets that connect task owners with specialized experts across a broad range of skills.||
|Contest-based labor markets||Public digital talent markets that source work through competitions in which multiple people perform a task or solve a problem, and a prize is awarded to the entrant who supplies the best results.||
|Specialized talent markets and services||Public crowd markets that are focused on a specific function, such as security testing, functional testing or data science.||
|Online collaboration platforms||Support mechanisms to facilitate discussion, sharing and hand-off of work products. Includes social collaboration, file-oriented collaboration and audio/video/screen share collaboration tools.||
Figure 3: Types of commercial workforce virtualization technologies
Experiments to build a toolkit and platform
At the Labs, we see commercial platforms as basic ingredients of a workforce virtualization toolkit. Many enterprises are already experimenting with these technologies in small ways. However, we have not yet seen a systemic approach to leveraging these ingredients to reimagine major portions of today’s work processes. This kind of comprehensive capability, supported by a full-fledged platform that brings new and existing ingredients together, is what we are pushing toward now. (See Figure 4). To build an Accenture platform, we are experimenting with various forms of crowdsourced workforce virtualization to identify what works well. Then we are developing experimental tools and techniques to mature our capability.
As part of our effort, we are adding functionality to the platform through the following projects, using a combination of commercial technologies and components that we needed to build out ourselves. (See Figure 5.)
|Crowd-powered data enrichment||Support workflows that combine micro-task crowd work with automation to enrich unstructured data into support business needs.|
|Self-service talent markets||Improve utilization and engagement of the workforce by facilitating best match between task and talent..|
|Crowdsourced software testing||Provide on-demand access to a large, geographically diverse, virtual testing capability.|
|Enterprise flash teams||Support rapid assembly of effective ad-hock teams from a combination of internal and external talent to accomplish high quality small projects quickly and cheaply.|
|Hybrid crowdsourced software development platform||Support a highly flexible, scalable software development capability that combines a smaller core team with workers drawn from internal and external crowds.|
Figure 5: Custom-built functionality in our Workforce Virtualization platform
In follow up blog posts, we will go into more detail about each of the projects, the motivations and the key research questions they address, and innovations that come out of them.