With Big Data 3.0, Marketers Will Finally Strike Gold
Until recently, Big Data has been a huge deal in marketing circles mostly because of its potential, rather than its present benefits. Like prospectors who occasionally find a hefty gold nugget nestled in a panful of gravel, analysts and marketers understand the potential value of the insights Big Data contains; the challenge is to extract actionable insights. The first iteration of Big Data—let's call it Big Data 1.0—didn't quite live up to its promise, and neither did the second iteration, Big Data 2.0. But the next generation—Big Data 3.0—is poised to finally deliver the mother lode marketers have long sought.
To put the evolution of Big Data into its proper context, it's important to recall that Big Data 1.0 came about because technology finally allowed us to capture a ridiculous amount of transactional or event-based data. But analysts quickly found that the dataset was overwhelming, primarily for two reasons: First, the scale was mind-boggling—imagine an Excel spreadsheet with millions if not billions of cells populated to get a sense of the challenges involved in parsing the numbers to yield meaningful insights.
And secondly, the complexity of the data was also hard to comprehend. In the pre-Big Data world, we grew accustomed to thinking about datasets as tables with columns and rows, numbers that could be expressed as graphs with an X and Y-axis. Encountering a dataset with multi-corollary information was disconcerting, to say the least. Moreover, it soon became clear that multiple relationships exist between patterns in the data, and that the sets go beyond two or three dimensions; they are multidimensional.
We were still perfecting the art of multidimensional analysis and visualization when Big Data 2.0 came about, driven largely by advances in natural language processing and sentiment analysis. Big Data 2.0 promised to be an improvement over its predecessor by featuring technologies that could detect emotion and sentiment in text—interpreting the feelings contained within tweets, posts and other text-based social media.
However, Big Data 2.0 has largely fallen short of its ambition. Detecting human emotion in text sounds great in theory, but it isn't easily done. That's because people tend to express themselves in complex and subtle ways. If you're having a bad day and someone asks how you're doing, it's unlikely that you'll respond with, "My sentiment today is negative and I am feeling frustrated."
However, keywords like "negative" and "frustrated" are what machines look for when identifying sentiment. To properly interpret emotion, technology has to take into account the delivery, including factors like tone of voice, cadence and intonation. Machines haven't figured out how to identify and match the appropriate "delivery" style to raw text.
Because of these shortfalls, both Big Data 1.0 and 2.0 have left a huge gap in the market—a chasm that marketers, ecommerce leaders and business leaders increasingly notice. Big Data technologies today are immensely powerful at capturing quantitative data—they excel in tallying clicks and similar metrics. However, bounce rates, site dwell time, click-through rates and similar indicators are fundamentally just numbers that reflect an activity. They yield metrics but are short on insights.
We predict that Big Data 3.0 will blend the two objectives of its predecessor versions—gleaning insights from both metrics and sentiment. But instead of driving more analytical value out of text, we predict that the market will turn to driving analytical value out of media that is currently ineffectively captured or explored: video, audio, surveys, blogs and shorter forms of text like tweets and Facebook posts.
How do we consolidate all of the market's quantitative and qualitative data into one dataset and run multidimensional analytics across it? Creating a platform that allows you to do that actually provides the context and insight you need to address the biggest problem in Big Data 1.0—understanding the "why" behind all of those numbers—and solves the riddle that has thus far eluded Big Data 2.0 by providing the emotional and behavioral context needed to understand the real sentiment behind what people are saying.
Imagine combining direct customer feedback captured from focus groups, in-the-wild video captured by market research, survey results and digital interactions (with websites and ads) into one unified dataset. Big Data 3.0 will see the sentiment and behavioral analytics developed in the 2.0 iteration applied to the dataset that Big Data 1.0 was never able to touch: non-text-based media. In this way, Big Data 3.0 will yield the rich marketing gold mine analysts have sought since the advent of Big Data.
Derek Carter is the co-founder and chief technology officer of Mountain View, Cali.-based YouEye, a customer research provider.