Crowdsourcing analytics: How to find the wisdom in the crowd

Crowdsourcing platforms use machine learning and advanced algorithms to identify who are the wise voices amongst the online crowd, and then turn that input into insights.

Image: iStock/oatawa

The din of the crowd can be confusing and even deafening. This holds true in public places, and it is equally true of big data coming from many and variegated sources. Yet, if marketers correctly perceive the messages that the crowd is conveying, it can make the difference in whether a product launch or promotion succeeds or fails.

How do you crowdsource analytics?

"What you want to find is the wisdom of the crowd," said Matt Bencke, founder and CEO of Spare5, which provides intelligent crowdsourcing software. "To get this wisdom, you need to find the right people in the crowd to listen to."

Just who those "right people" are depends upon the business problem to be solved. In everyday life, most of us perform this task through inquiry. For instance, if we are new to an area, we ask who are the best doctors, the most reputable auto mechanics, and the finest accountants.

Companies like Spare5 apply this principle to the web with an intelligent crowdsourcing platform that uses machine learning and advanced algorithms to analyze the din of the online "crowd," determine who the wise voices in the crowd are, and then turn the input from these sources into actionable insights to companies. Finding these insights, and focusing on the best sources for information, can be invaluable for organizations that are struggling to make sense out of mountains of audio, video, and unstructured text coming at them from all directions.

"We make it our job to curate data for customers, once we understand the business problems that they want to solve," said Bencke. "Let's say, for example, that an organization wants to better understand a certain demographic such as women and their attitudes toward fashion. A combination of machine learning and algorithmic intelligence might identify the premier spokespersons from this demographic by finding women in the group who are in some way associated with fashion."

"What the software fundamentally does is predict the probability of correctness with the human being most likely to give the best answer to a question," said Andy Ganse, Spare5 data scientist. "From there, the machine learning will further refine the algorithmic approach."

SEE: Executive's guide to Big Data strategies and best practices (free ebook)

In addition, crowdsourcing analytics technology can enable companies to better tune their approaches to market studies.

Bencke gave the example of a retailer that had two pricing levels for an item--either an $8 per unit price point for persons who didn't want to spend too much, or a per unit price range starting at $100 for premium buyers. Unfortunately, there were no product price points between these two extremes for a mid-market crowd. By interviewing the "crowd" and identifying voices in the crowd that were astute about customer tastes and pricing sensitivities, the retailer was in a better position to scale its pricing and to target segments of the middle market that it was missing.

"There are still many companies that are largely unaware of the value of crowdsourcing analytics, and how to derive intelligent insights from it," said Bencke. "They have tried to do this in-house, but often find that they lack resources. However, the thinking about this is starting to change. Companies are beginning to understand that, instead of trying to make sense of the collective shouting in a stadium, you can employ intelligent tools that enable you to discover that the voice in the stadium that you really want to hear belongs to a person who is sitting in section 124. Machine learning, intelligence, and probing can help you to achieve that."

Also see