Charles W. Eliot, the former president of Harvard University, once noted, “All business proceeds on beliefs, or judgments of probabilities, and not on certainties.” So, since according to Eliot there is no certainty in business, decision makers need to rely on guesswork or probabilities, and of the two, analytics and probabilities will surely give better results.
Big data promises to provide insights that traditional data and analytics can’t achieve, but many companies become so excited about the idea of finding a hidden nugget that nobody else knows that they forget to focus on a potential solution to a particular business issue. The challenge is identifying whether big data is sufficient by itself to generate the desired insight, or whether existing data or combined data could achieve comparable results that lead to better decisions.
Two recent use cases demonstrate the value of big data in providing insight that traditional data alone could never achieve, and they’re likely to become the “killer applications” of big data.
1. Quantitative trend spotting & early warning systems
Recent research published in the Journal of Marketing Research has shown that using big data from Google Insights for Search along with syndicated data and POS/scanner data can identify emerging trends in asthma drugs, grocery items, and wine, among other products. This insight enabled analysts to quickly identify trends without having any specific domain expertise or input from professional trend spotters — an incredible advancement in trend identification.
2. Marketing mix efficiency analysis
Other researchers combined traditional data and big data to show the value of social media marketing, a notoriously slippery metric. They showed that 25 percent of the sales variance over the amount explainable by advertising and media expenditures could be directly attributed to company-facilitated Facebook conversations.
Traditional marketing mix efficiency analysis is based on the relationship between sales and marketing spend, but big data helps identify other variables such as daily or weekly searches, social media mentions, and requests for information that the company can correlate to weekly sales and marketing expenditures.
Using these additional variables helps explain formerly unexplained model variances and provides more insight that the company, and Chief Marketing Officers, can use to help demonstrate that marketing efforts are influencing consumer behavior and how they may lead to sales over the longer term.
Big data and these killer applications are providing probabilities and insights that are coming much closer to the formerly unattainable certainties that Eliot alluded to, and business is better as a result.