Crowdsourcing: New Trends and Developments (Part 2)
In Part 1 of this two-part series, we provided an overview of crowdsourcing, defining what it is and how it is different from online freelancer marketplaces. We not only provided examples of different crowdsourcing platform providers (of which there are many) but also provided illustrations of real crowdsourcing in action.
Today in Part 2 of this series, we cover the emergence of practices and functions to effectively manage crowdsourcing across organizations and some of the segments where crowdsourcing has grown both on the demand (buyer organizations) and supply (platform providers) side. We also look at how the space is evolving and provide some highlights and suggestions for practitioners.
Large Organizations are Starting to Manage Crowdsourcing
Competencies, practices and models for harnessing and managing crowdsourcing within and across large organizations are beginning to take shape. A prime example of this is GE GeniusLink.
GE was an early adopter of crowdsourcing for its own “open innovation” initiatives. Over the past few years, GE evolved the GeniusLink organization, which has established ongoing relationships with more than 20 online platforms. GeniusLink not only acts like an “internal MSP” for managers within GE who want to engage in crowdsourcing, it has now also become an profit center by selling its services to other businesses outside of GE.
In another example, professional services firm Deloitte launched Pixel in 2016. Like GE GeniusLink, Deloitte established ongoing relationships with a range of platform providers. And while Deloitte created Pixel to serve clients who had problems that could be addressed through crowdsourcing, it was also a complete solution with a methodology and professional services to bring clients from point A to point B.
Finally, at more or less the same time, the US. Federal government has also established crowdsourcing and citizen science online resources to promote and guide the use of crowdsourcing across agencies.
Some Niche Categories are Filling Out and Growing
As noted above, crowdsourcing takes many forms and can accomplish many different kinds of tasks or projects. Looking across the many forms and applications of crowdsourcing (only some of which are mentioned in this brief), one almost gets the sense that, at this stage, we are really just scratching the surface. At the same time, some niches are beginning to take shape, where the application of the crowdsourcing model provides a unique solution for an otherwise intractable problem. For example, there is a niche of software testing for distributed applications, localization or cybersecurity, where there are already a number of “crowdtesting” platform providers:
- Crowdsourced Testing
- Global App Testing
- Test IO
- MyCrowd QA
Other crowdsourcing niches that seem to be filling out and growing are marketing/advertising related and software and algorithm development related.
Also, as dystopian as it may sound, using humans to perform simple microtasks in order to train machine learning and enhance artificial intelligence systems is a big and growing business. There are already a number of platforms specialized in this area, including Mighty AI, Figure Eight, Scale and Playment. (Note: The new branding of CrowdFlower to Figure Eight (4/2/18) reflects the company’s strategic shift to focusing solely on “human-in-the-loop” AI training.)
At the same time, it is important to bear in mind that we are likely just starting up what looks like a long learning curve over which new niche applications will be discovered. For example, crowdsourcing is finding applications close to home, in the area of procurement. Examples of this would be Beroe’s LiVE Poll category management/spend analysis crowdsourcing solution, Coupa’s Prescriptive Community Intelligence and a recently launched startup called Procurement League.
All of what we discussed above represents just a slice of the whole crowdsourcing pie today. One thing is for sure: crowdsourcing use is not only increasing, but how it can be used and what form can take is also changing.
Amazon Mechanical Turk and Innocentive, arguably the progenitors of the current crowdsourcing space, were founded over 10 years ago. From that time until now, obviously a lot has been changing, including, over the past several years, the more widespread adoption of crowdsourcing by larger organizations and discovery and diversification of applied uses. But beyond that, crowdsourcing technology and models are also beginning to change.
We mentioned above that a growing number of platforms have focused on using a crowd to train AI and machine learning (ML). But now some platforms have gone a step further combining AI/ML and human activities to produce various kinds of services (e.g., Directly delivers customer service solutions).
Not surprisingly, starting in 2017, we have begun to see the arrival of new blockchain-based crowdsourcing platforms, such as:
- Gems (bills itself as the Decentralized Mechanical Turk, powered by Ethereum)
- STORM (enables microtasking(gamefied) and efficient micropayments worldwide)
- StartCrowd (a challenge and collaboration network facilitating AI projects)
- Effect.AI (an open, decentralized network that provides services in the Artificial Intelligence market)
Blockchain, distributed ledgers and even cryptocurrencies may pair well with crowdsourcing, which is an inherently a decentralized model. Blockchain also natively supports smart contracts and is very well suited for micropayments.
Much has happened since our first brief on this subject almost three years ago. Crowdsourcing has certainly been proving itself to be a valuable, innovative tool kit for business problem solving within large organizations. And some organizations have been establishing ways to promote, facilitate and manage an entire organization’s use of crowdsourcing (e.g., GE GeniusLink). At the same time, while enterprises increasingly get their arms around it, crowdsourcing continues to evolve. New applications continue to arise (a good thing), while the crowdsourcing technology-based platform remains necessarily dynamic.
In any event, crowdsourcing is not going away. On the contrary, it is likely to become woven into the fabric of most large organizations over time and increasingly become an important sourcing alternative for many types of traditional or new services. For example, if an organization is increasingly using AI/machine learning, it may need a crowdsourcing provider like Mighty AI or Figure Eight to normalize data and train the AI. Or if an organization needs to find a solution to a complicated problem (e.g., a new sweetener formula, a prediction algorithm), it may want to launch a crowd challenge through through Innocentive, HeroX, Kaggle or other providers. If it does take that approach, it might pay $1 million for the solution, but it will avoid the possibility of investing likely more than that amount in traditional R&D and not end up with a solution.
Crowdsourcing platform providers can deliver solutions that can be beneficial to an organization in various ways:
- No contingent worker co-employment risk, since service delivery (outcomes, results) comes from a transient crowd and is contracted (SOW/MSA) with an incorporated provider which is accountable for delivering
- Tends to be cost-competitive with more traditional alternatives (at least for certain use cases)
- Payment occurs only when the desired result is achieved (i.e., in contests, challenges)
- Crowdsourcing can even deliver solutions that would not otherwise be achieved
While crowdsourcing providers are in many ways similar to all other services providers, they are in many ways quite different. For example, in the extreme, they may be blockchain-based platforms that pay micro-task workers with tokens/cryptocurrency in any part of the world. Also, using crowdsourcing may require change management in an organization (something that GE learned to address), and it will certainly require that procurement groups, at the very least, review their assumptions and practices in places for dealing with suppliers of services.
In any event, how an organization manages crowdsourcing is a question that will increasingly require answers. We will return to this and other questions in future installments.
But in the meantime, may we suggest crowdsourcing?