February 22, 2016
Putting Crowdsourcing to Work for the Enterprise (Part 4)
By: Alex Kass

Digital talent brokering

Through this series of posts we proposed a novel model of crowdsourcing for enterprises that we call “workforce virtualization.” During the past two years at Accenture Technology Labs, we have been studying, designing and building a comprehensive suite of supporting technologies and methodologies under the umbrella of a workforce virtualization platform. We believe that this platform will be the cornerstone to help businesses adopt and industrialize this new model of work across the enterprise.

A core conceptual and technological underpinning of workforce virtualization is the idea of "talent brokering"—that is, how to enable task owners and talent owners to find each other and to create a match between a given task and the available talents based on some combination of desired parameters (e.g., skills, languages, turn-around time, location requirements). In this post, we discuss a talent-brokering platform that we designed, built and piloted internally called Digital Talent Broker. This platform has been serving Accenture globally across 20 geographical locations. It has more than 9,500 users and has brokered over 13,000 hours of work, with zero transactional cost.

Why internal talent brokering?

Today, we are seeing a proliferation of vendor talent brokering services that offer contingent labor to augment or replace traditional workforces. Examples include UpWork (freelancing), Taskrabbit (labor), TopCoder (context-based application development) and Amazon Mechanical Turk (microtasking), just to name a few. Enterprises have also started to experiment with these platforms, typically leveraging the on-demand labor to augment parts of a business process, often involving tasks that are high in volume and low in complexity (e.g., making image judgments for moderating social networking platforms).

However, there’s an open question: Can digital labor markets and talent brokering be used to systematically transform the way tasks of all sizes, durations and complexities are allocated, as well as how talents are managed internally within an enterprise?  We created Digital Talent Broker to explore and answer this feasibility question.

Digital Talent Broker is a web-based platform with a well-defined workflow for creating and brokering work between job posters and job candidates. Operational scalability is one of the key tenets, including:

  • Self-service—Tasks owners can broker work with talent owners without involving an intermediary. This enables task owners to find the right talent and gives talent owners the opportunity to seek tasks that match their interests and/or skills.

  • Automation—Tasks and talent owners are automatically matched based on desired parameters such as skills. This not only creates more suitable task-talent pairings, but also reduces the chance of communication overload.

The brokering process follows five stages of interaction:

  1. Post a job—A job poster creates a job post that includes details such as type of job, summary, estimated duration, start/end dates and desired skills. The job poster also creates a desired candidate profile such as location and level of individual. Emails are proactively sent based on the job profile.

  2. Discover and apply for jobs—Candidates find open jobs and apply if interested. Discovery generally happens either through a targeted email or through browsing the list of open jobs.

  3. Make an offer—The job poster reviews the applicants and selects the best candidate. The job poster is provided with links to each candidate’s internal resume and social profile page.

  4. Accept an offer—The candidate has the option to accept or reject the offer. Accepting the offer completes the negotiation, while rejecting it prompts the job poster to select another candidate.

  5. Exchange feedback—Once the job is complete, the project supervisor can provide feedback to the candidate through the system.

Job Jar deployment

Our first deployment of Digital Talent Broker was for Accenture’s North America Job Jar program—an internal contingent labor program created to increase the utility of unstaffed employees. Originally Job Jar was administered by HR specialists who brokered tasks manually, by routing communications back and forth between the job requesters and available unstaffed employees. Although the program was moderately successful, three challenges limited its scalability:

  • High transaction cost—Each and every task brokered had to be manually identified, vetted and communicated with candidates.

  • Brokering lag—All communications were funneled through HR specialists, which created a bottleneck and delay.

  • Mismatch—Given the reasonably large number of candidates and the lack of support for automated matching, HR specialists often reached out only to people they knew.

We piloted Digital Talent Broker in two geographies in North America in order to transform the process into a self-service, market-driven model. Interest and usage has grown steadily. As mentioned above and shown in Figure 1, Digital Talent Broker for Job Jar has been used by more than 9,700 employees across 20 geographies. Over 500 jobs of all durations and complexities have been posted, with each job averaging five applicants. Overall 13,000 hours of work have been brokered through the platform.

Fig. 1 Digital Talent Broker for Job Jar Growth

Figure 1: Digital Talent Broker for Job Jar growth

Adding additional talent markets for more advanced skills

The early success of Digital Talent Broker generated interest in broadening the use of the approach, and has since led to several implementations and extensions beyond Job Jar. For example, we deployed specialized talent marketplaces for solution architects. Creating solution plans often involves tapping into the expertise of several specialized domain experts, who may be widely dispersed within the company. To support this expertise sourcing, we added a profile creation page to Digital Talent Broker that allows solution architects to define their skill sets using a standard taxonomy. In addition, we extended this version of Digital Talent Broker to target notifications for jobs to talent that matches the skill required. This digital talent market is now in being used to broker a steadily growing set of tasks.

Next steps

Based on the success of these pilots, we are continuing our work in two areas:

  • Contributor’s genome—Creating a deeper understanding of the skill sets of members of the crowd and a platform to automatically update this profile based on training and career progression.

  • Talent brokering as a service—We believe the future workforce will be a blend of long-term employees plus contingent labor accessed through vendor platforms. In order to optimize the allocation of work among these labor pools, we have created connectors from Digital Talent Broker to external labor pools that enable a single job description to be targeted to multiple sourcing channels. We are also creating a service layer on top of Digital Talent Broker to enable crowdsourced software development and testing to leverage talent brokering as a component.

For more information about crowdsourced workforce virtualization, please contact Alex Kass at Thanks to David Q. Sun for contributing to this post.

Industry & topics highlighted


Popular Tags

    More blogs on this topic