Most tools in the booming field of analytics rely on quantitative data. Analytics tends to work best with large numbers of numerical data points; smaller data sets and those where the data isn't easily quantified can present problems.
It's with those trickier data sets that YouEye is betting it can help. On Tuesday, the startup launched its Experience Analytics Platform, a tool designed to capture and structure qualitative and unstructured data, including video, and then give users the power to run analytics on it.
Using YouEye's platform, enterprises can measure reactions to marketing campaigns, mobile-app designs, brand positioning and sales funnels, for example. To enable that, YouEye conducts online studies with volunteer participants, who first have to download its testing app. Participants go through the test sessions, being recorded with video and audio. Their on-screen activity is recorded, too.
Five main components in YouEye's Experience Analytics Platform then help turn that data into actionable insights.
The Behavior Metrics module taps a combination of trained human analysts and machine-learning technology to analyze video for key sentiments expressed by consumers during a testing experience, including frustration, confusion, praise and recommendations. Those measures are then quantified for further analysis.
A Highlights Creation component condenses raw video data into highlight reels that capture the most important moments of the test session. A Survey Results component integrates survey data into video data, while a Research Tasks module guides researchers as they analyze results, highlighting key patterns of behavior or sentiment or suggesting areas for further analysis.
Finally, a detailed Demographic Segmentation component allows users to combine their entire dataset with rich demographic segmentation data for a more nuanced understanding of the results.
With the data in place, users can analyze it for patterns and insights. A metric that has proved useful for clients focused on e-commerce, for example, has been the placement of a discount code during the shopping process and how that affects shopping-cart abandonment rates, said Collin Sebastian, YouEye's vice president of product and marketing.
YouEyes pricing typically ranges from $8,000 to $15,000 per testing assignment, depending on how many participants are involved; typically that group size ranges from 25 up to several hundred people. Google, Facebook, Walmart, Salesforce.com, eBay and Etsy are among the companies that have used YouEye's services.
Making qualitative data pay depends on what you use to analyze it, including natural-language processing engines and machine-learning algorithms as well as "hours of training in your industry," said Ray Wang, founder and principal analyst with Constellation Research.
"Companies like IBM, Clarabridge and Lexalytics have been doing this for customer service and experience for some time," Wang said. "Its not easy, but the technologies keep getting better."
YouEye's technology represents the "next generation," he said.
Of course, "qualitative data" can mean different things to different people. Tom Anderson, CEO of text-analytics company OdinText, defines it by the number of data points, or subjects, rather than the unstructured nature of the data itself.
"You cannot have patterns in what five people say, because you can't apply statistics or math," he said.
The question is, how many data points do you need before analytics start to make sense?
"We tell people there are some things you can get out of small sets of qualitative data, but honestly, it's not much," Anderson said. "It's not something you're going to base business decisions on."