One of the most inspiring parts of being a VC is helping build really big ideas that change the world.
Incredibly interesting societal changes are underway today arising from the new ways we communicate en mass thanks to the ubiquity of the Internet. 39% of the global population, or 2.8Bn people are connected today. The next billion will join in less than five years, at which time fully half of our population will be connected, importantly with most of that growth from the developing world thanks to the mobile Internet. We have reached a critical mass of global connectedness.
Adam Smith’s division of labour has generated every increasing returns commensurate with increasing specialization of task assignment. Crowdsourcing, and crowd labour in particular has been looked to as the logical extension to drive these increases in productivity through the 21st century. However, my view is that we’ve been climbing toward the peak of inflated expectations on the gartner hype cycle for some years now and for many crowdsourcing is now falling toward the trough of disillusionment. This is a time for investors to apply lessons learned rather than simply “shoot for the moon”.
As we’ve learned from the adoption of other technologies, there are usually leading applications of an emerging technology which foretell its usefulness and inform effective exploitation. In the case of crowd labour, there have been hundreds of experiments running both on Amazon’s MTurk and through a plethora of startups. Most of these have not met expectations.
While it will be some time before, “economies are transformed” by crowd labour, there are bright spots worth understanding. Specifically, we’re seeing that where existing large scale business processes are today being executed with captive labour pools it can be highly effective to implement software that manages this workflow. Our recent investment, crowdcomputing.com, has implemented such software in a way which then blends in crowd labour — with effective quality management — to scale that captive labour force, and simultaneously applies machine learning to start automating tasks which don’t require human intelligence. This approach of starting with what works and extending it rather than “starting over from scratch” is proving effective.
In particular, CCS’s approach seems to work well in markets which deal with ongoing streams of data rather than having “one-off” custom jobs to deal with. For example, Financial Information Services (FIS) providers are characterizes by having such workflows to sustain the products they offer their customers. FIS vendors were early adopters of business process outsourcing (BPO), which now has well understood capabilities and limits within those organizations. Bringing crowd workforces to bare has many parallels.
Thinking about human labour as a programatic element within a software managed system is at first Orwellian but the resulting opportunity to reduce the assignment of human talent to menial tasks better performed by machine learned engines is perhaps the opposite.
A powerful yet unrealized promise of the Internet is bringing on-demand, scalable labour to bare on tasks today performed by dedicated workforces. We’ve seen a glimpse of the compelling economics and flexibility that can be achieved with first generation crowd labour. CCS is taking that essential next step of wrapping crowd labour within a machine learning infrastructure that can assure quality and optimize resource utilization to achieve existing, real-world business objectives. That’s a really big idea that can change the world.