Got Social Data? Must Transform Analytics Operations

Data is nothing new to marketing and business operations. However, it’s come to the forefront of marketing strategy because marketers are now increasingly the stewards for business use of data. No longer does the data live in some deep, dark, remote location where only a few people have access. Marketing departments rightly own the data operation because business gets done via a large, interconnected, and matrixed series of customer connections – digital, offline, emotional – and marketing is an owner of the customer relationship.

Connections are powerful. We buy from companies (both as consumers and business professionals) when we have a connection to the brand and the people who work there. Could be the sales or account manager at your automation vendor; the single sign-on and coupon-free shopping you have at an e-commerce site; the one-click access to information through a mobile app; or the barista at the local coffee shop. The need is for marketers to connect customers with experiences and purchase opportunities – and then get the heck out of the way.

Marketers create those experiences through the responsible use of data. Now that we have so much data in so many forms from so many sources, the data management question gets thornier. Automation technology certainly helps us act on data, but the data has to be there – accurate, clean, timely, and analyzed into a sensible set of actionable insights. Structured data – from demographics, customer forms and purchases, and email response rates – is pretty straightforward to analyze and use. The unstructured (or multi-structured) data is increasingly the challenge.

I find that too many marketers realize how hard it is to incorporate social data in analytics, and simply stop. Either they don’t use the data at all, or they use it in a silo that is disconnected from the rest of the organization (and thus hard to fund and make more meaningful). Often, the business objectives and the social marketing objectives are not in alignment. For example, the business objective is to attract mature buyers of fine wool products, but the social strategy is creating a conversation on urban youth fashion. The data and contacts from the social conversation are not helpful for connecting with primary buyers.

While social marketing has value in many ways to many marketing organizations, the real point of having a social analytics program is to understand the sentiment of your customers and marketplace. Use that sentiment to reveal opportunity or to uncover future vulnerability. As a result, marketers are scrambling to measure social marketing and incorporate it into the larger marketing-driven program.

For example, Cisco uses social media to monitor chatter from both customers and channel partners about suggestions, trends, and problems. “We return that insight to the business. And when you respond to let customers know you’re listening, it can go viral and make a big impact on the business,” says Sherri Liebo, VP of segment marketing and marketing communications at Cisco Systems, in a panel at the recent NCDM big data conference. “I always encourage marketers to consider what business cares about.”

Cisco uses share of voice, “as well as the tonality of that share of voice,” as the primary social metrics, Liebo said at the conference.

Shiv Singh, global head of digital at PepsiCo Americas Beverages, said in the same panel session that social media in support of business is about tracking brand health. “When we know how our brand is doing, we know how that will translate directly into sales,” Singh said. PepsiCo – with its retail operation as well as a significant B2B reseller network – also uses social media analysis to track competitors, sometimes unearthing rivals’ actions before they’re announced, Singh said.

That is pretty powerful stuff! However, the curse of multi-structured data is the variety in data types and degree of structure. Many methods and algorithms that have been developed for structured data (the kind that fits neatly in rows and columns) may need adjustment to accommodate for these social and clickstream data streams. Segmentations – a key output of most data modeling – can become sparse and harder to group when modeling multi-dimensional data. However, the kinds of data characteristics that are typically employed for marketing analytics may still apply – things like sequence (time and symbolic (unplanned)), graphical, spatial, and location-based and multi-media (web). Expect new ideas will emerge in handling this data so that our existing algorithms can adapt to handle it.

At the end of the day, what we need to do is figure out (and focus on) the business problem we are trying to solve. Are we looking for insight on where to put more capacity for our operations or data center? Are we looking for the next product to launch? Are we looking for new markets to enter? Are we looking to validate the value of our current product mix? The management of data must always be in service to the business objectives. Marketers can start by asking the right questions of the data analytics folks, and then start to incorporate social sentiment into those equations.

How are you using social data today as part of your analytics program? Please share ideas in the comments section below.

Social Data image on home page via Shutterstock.

Related reading