Advanced analytics is driving many of the greatest operational changes in retail.
Analyzing data and leveraging the precision that comes from detailed knowledge of transactions has revolutionized practices such as inventory management, fulfillment, e-commerce and in-store marketing.
But the biggest paybacks from advanced analytics remain largely untapped and are just now emerging: data mining of unstructured data in the most robust area of consumer communication and interaction—social media.
"We are not there yet. We have robust customer data on physical and e-commerce. But social media data mining is an area of the strategic roadmap that still needs to be done," says Michael Ross, vice president for digital shopping and customer marketing at Meijer, the Grand Rapids, Mich.-based operator of 229 stores.
"It is high on our interest level. It is a key imperative for us. We want to connect household data with social media activity. If I had to predict, we will be there within 18-24 months. Then we'll be able to leverage social media for business paybacks at a much different level," Ross says.
Ross is not alone. Much has been written about the promise of data mining, and certainly social media has radically changed how all shoppers communicate, learn, share, form opinions and set trends.
But mining that diverse range of data and customer sentiment remains elusive. The tools to deeply mine unstructured data, especially from social media, are just now maturing and coming into the market. Retailers continue to struggle with integrating disparate silos of information, which is key to data-mining success.
"I have talked to a couple of companies that are developing the tools to be able to leverage social media data now. It [also] all ties back to having a single view of the customer," Ross notes.
In the era of social media, understanding what consumers are wanting, feeling and saying, as well as where they are shopping and spending, is paramount.
"The industry started out measuring social media impact but were using the wrong [key performance indicators]. First they were reading the 'likes' and then 'shares.' The problem there is that measuring those KPIs does nothing to put money into anyone's pocket," says Teri Ross, president of Imagine That Consulting Group dba Your CMTO, St. Paul, Minn.
"The next buzzword was engagement and replying to what you posted. This of course generates buzz, but still leaves the question of how to track back to revenue. Hello, isn't that why we are doing this?" she states.
"Social media data mining is an area of the strategic roadmap that still needs to be done."
COPING WITH NUMBERS
The challenges to leveraging social media for actionable insights leading to profits is formidable, she says. One major challenge in mining social media activity is the sheer number of devices, including laptops, smartphones, desktops, tablets and others.
"Also, the number of social platforms has proliferated. It is not just Facebook and Twitter. We now also have SnapChat, Pinterest, and so many others. The number keeps growing," she says.
According to Teri Ross, the answer to "mining" social media today involves "data scientists" who can pull data from myriad and disparate sources and then leverage existing tools to identify profitable trends.
"No one today is going to be successful if they are not mining social media. But there is no one tool. That is why the number of data analysts, who can utilize existing data analytics tools, is growing so quickly," Teri Ross says. "Pulling all the data together is key. A whole new level of expertise is required: data scientists."
Leading analytics vendors are investing massively in a race to bring to market data-mining solutions built on machine learning, natural language processing, and other highly advanced technologies.
According to a new report by IHL Consulting Group, the market for retail analytics, including data mining, will grow $4.6 billion in 2016 to $7.1 billion in 2020, an increase of nearly 54 percent.
ACCESS TO ANALYTICS
"As retailers move towards unified commerce, the access to and quality of specific analytics will be critical to increasing business," the report states.
Emerging data-mining tools are able to analyze, identify and predict shopper behavior and non-intuitive, yet highly profitable, retail opportunities.
But so far, that level of data mining remains in an early stage of rollout vs. a mature reality. For example, Christopher Holloman, chief data analyst at Information Control Company (ICC), a data analytics firm, pointed to the growth of customer sentiment analysis as well as other areas of advancement.
"Most of what I am seeing is people pulling texts, tweets, or Facebook posts in high volume and then starting to do some natural language processing to pull out [customer] sentiment," Holloman says.
Retailers are seeking to understand sentiment by age group, geographic location, and similar types of attributes. "But beyond that it is the question of how does sentiment work to drive sales," he adds.
So much of social media has been the push of information rather than the pull—so far, says Leigh Helsel, principal of Retail at ICC: "Certainly there is more work to do."
Leigh noted that companies are increasingly monitoring social media about their competitors and seeking to reap strategic intelligence in that way.
"What are people saying about all-day breakfast, or Burger King launching the Mac and Cheetos, or limited-time offers? How are they resonating? Is there a lot of chatter or little chatter? And then trying to mine information around that," Leigh says.
In grocery, "we know that that Columbus, Ohio, has been a test market for BOPUS [buy online, pick up in store] for Giant Eagle and Kroger. They are looking at social media to gauge what customer sentiment is," Leigh says. "What are people saying about their experiences with Giant Eagle? Retailers are looking at that with regard to new product offerings or services" and what people are thinking and feeling, she notes.
A further emerging key for retailers and vendors is analyzing social media to determine whom people are networking with and who the most influential trend setters are, Holloman adds.
Decision-makers are becoming more and more interested not just in what people are tweeting but rather "to whom people are tweeting and identifying the tastemakers in terms of buying power and influence," he says.
There are more tools available than people are using. Much of the focus now is on how to get better information out of the networks themselves, Holloman notes.