Alexander Nicholas Kozlov
Soliciting ideas from large numbers of people through crowdsourcing can be a great and cost-effective way to solve problems, generate ideas or gather opinions. Companies like Starbucks, Samsung, Lego and Airbnb have successfully used crowdsourcing models to test new products, engage with customers and build brands.
Many businesses use crowdsourcing models internally to solicit ideas for innovation. The rationale is that the people directly involved in interacting with customers, making products and processing information can offer unique insights into how to add value and improve operations.
For a large, global organization, the challenge in applying this suggestion box approach lies in assessing potentially thousands of suggestions. How do you find the truly innovative ideas and separate them from the also-rans – and do so in a timely manner? How do you ensure great ideas don’t get lost in the shuffle? Some businesses have used online idea generation platforms that incorporate gaming techniques to encourage participation and communication. Real-time translation systems can help global companies with multi-lingual teams.
Nonetheless, quickly assessing ideas to identify the standouts can be difficult. Today, advances in machine learning applications and cognitive analytics tools are helping organizations dramatically improve their ability to analyze crowdsourced innovation proposals and identify the ideas that have the most potential. The tools assess proposals based on a variety of criteria, including how often they are shared and how frequently they’re commented on. They also analyze keywords to gauge sentiment – i.e., a comment like “This is amazing!” from a peer would likely be tagged as promising. Over time, the tools continually improve by assessing past performance for lessons learned that can be applied in the future.
One of the intriguing aspects of these tools is that the logical path by which learning algorithms reach a particular conclusion often isn’t clear and can’t be traced back to specific rules. Indeed, the intricacies of machine-based logic is becoming an increasingly hot research topic.
My colleague Ricardo Garza recently wrote an article in Cognitive Business News that discusses crowdsourcing innovation and machine learning in greater detail. In this piece, Ricardo also shares how Softtek helped a large North American grocery chain apply machine learning to crowdsource innovative approaches to engaging with customers.
