Big data is altering the rules of marketing.
The Web is helping shoppers find what they want, when they want, on the devices they want to use, and thanks to big data and predictive analytics, shoppers’ wants can find them, too. Retailers now know when a TV ad inspired a mobile search, or when a retargeted banner produced a click-through and purchase—and they can apply that information to future interactions.
For direct response marketers, it’s no longer just TV and telephone; there’s desktop, tablet, email, mobile, apps, and a few dozen other touchpoints that cleave off from there. Just 10 years ago, the biggest data showed credit card activity, and the only shopping algorithm in evidence powered Amazon’s simple suggestions. Now, data tracks every preference along the purchase path and into the backend.
“The amount of data out there—not to mention the number of media channels—is growing by the second, and marketers are desperately scrambling to keep up,” says Monica C. Smith, CEO and founder of Cedar Knolls, N.J.-based i.Predictus and Marketsmith, Inc.
But marketers that can analyze the billions of bits of information consumers leave behind as they research and browse products online or in-store have a big advantage over companies relying on traditional methods alone. And the secret—although it may seem counterintuitive to many companies—is to share that data throughout the enterprise.
The Opportunity in Data
Big data can help enhance effectiveness in nearly every aspect of retail operations. Retailers are adopting big-data solutions that promise to reduce fraud, increase customer loyalty, streamline operations, boost sales, and more. But the chief concern for DR marketers is to ensure that their ad dollars are producing awareness, sales, and ROI.
“Until recently, DR marketers always focused on a single question about channel effectiveness: ‘How much revenue did a specific media channel generate?’” Smith says. “Now, while the game is still the same, the proliferation of data has created a new challenge. DR marketers have to find new ways to measure media effectiveness.”
While attribution will never again be as simple as a dedicated 800 number, new algorithms and dashboards from i.Predictus and other experts can gauge the effectiveness of individual airings by incorporating data related to touchpoints such as website visits, click-throughs, and call center interactions. Marketers can still judge product or campaign success based solely on profits and losses, of course, or they can collect data and use it to corral their success into additional sales.
“Maybe one channel [or] message drives a lot of traffic,” Craig Handley, CEO and co-founder of Listen Up Español, offers as an example. “The channel driving lots of eyeballs could be very costly, while the one driving fewer eyeballs could be very profitable. Maybe the channel driving the most eyeballs just needs a tweak in the message to become profitable. The summary here is that data is what drives ROI.”
Lending Context to Information
Beyond boosting campaign efficiency, big data permits increasingly exacting targeting, or “contextual” marketing, that takes into account the consumer’s browsing history, interest, frame of mind, location, and other factors that can affect their willingness to buy at a specific moment in time, even incorporating the context of weather and traffic. Multiple channels support the customer’s journey through the purchase funnel, and no two journeys are exactly alike.
“There are even more opportunities in modeling cross-media attribution,” Smith says. “How do multiple advertising exposures across multiple channels drive a consumer’s journey—or alternatively, inhibit his or her journey? Centralizing all advertising data within one repository can open up more complex modeling strategies to brand marketers. The sky’s the limit.”
I don’t believe in ‘misleading’ data, but there’s always the threat of incorrect data, irresponsible statistics, and weak insights.
The explosion in mobile usage has added a new level of complexity and knowledge, she notes. “Mobile has upped the ante. Like real estate, it’s about location, location, location. Marketers can now use location, provided by a consumer’s mobile device, as a contextual marketing element. Combined with product-interest patterns, the result is a contextual pattern that is more predictive than any single variable.”
Although it’s often considered only a brand-building medium, TV has also upped its game, data-wise. “We see cable providers that are able to target extremely customized segments of audiences based on data from Experian and Transunion,” says Andrew Blickstein, founder of Home Run Media in Chicago. “Machine-learning techniques are able to capitalize on viewership data and better understand patterns of viewership and response that allow advertisements to feel more individualized.”
As more retailers gain access to such data, it will become even more crucial for marketers to anticipate shopper preferences. “It becomes not only possible but essential to provide offerings designed around consumers’ real-time needs,” says Tina Wisner, vice president of data science and analytics for Winter Park, Fla.-based DR agency Kre8 Media. “To uncover the consumer’s moment-by-moment sentiments and behaviors that drive action through wholistic data is going to be the differentiator.
“When we talk about attribution, we sometimes look at it as who gets the credit for a particular customer, but it isn’t really who gets the credit. It is how many touchpoints brought this person in, and what combination of touchpoints, and in what sequence,” Wisner says. “From a business standpoint, it’s understanding the interaction between the different media that brings this consumer on board.”
Defining the Data
The problem many marketers have with big data is that it informs so many parts of the process. Of the retailers surveyed by Internet marketing consultant Monetate last year, most (62 percent) said they wanted to focus big-data initiatives on merchandising; they also hoped to use data to streamline operations and enhance e-commerce. But those same respondents said they expected big data to affect marketing (30 percent) and e-commerce (20 percent) first. “Data—large or small—should expose a lot of areas that [an] organization should be able to optimize,” Wisner says.
If a company is getting quality leads but not conversions, for example, it may expose an issue such as a need for more sales training or improved call-center scripts. “That’s a simple example of how marketing data exposes other issues in the operational structure of the organization,” she says. “The media sometimes can help you identify pockets that are converting well, but it is really what we discover looking at the trends.”
Marketers must focus on what they want to know, and define the indicators that are going to tell them. If big data sends a campaign off in the wrong direction, it won’t be the data’s fault. “Data doesn’t lie,” Smith says. “It’s raw material. How it’s interpreted is the key to creating good, actionable decisions and also where mistakes can be made. The best action is to start with key, measurable performance indicators.”
“It’s important to carefully think through how each piece of data will be put to use,” Home Run Media’s Blickstein adds. “Every time we contemplate a new advertising medium, we have to sit down and strategize how and what data we will need to collect, how we can go about collecting it, and what reporting will look like. I don’t believe in ‘misleading’ data, but there’s always the threat of incorrect data, irresponsible statistics, and weak insights.”
Integrating big data with traditional data and making sure that the information is clean and actionable is much more difficult. “The biggest challenge, in my opinion, lies in the expectation that we need to find added value from these new data sets,” Kre8 Media’s Wisner says. “New metrics must be defined, and must be converted into a format that reveals hidden patterns and correlations.”
Mobile has upped the ante. Marketers can now use location, provided by a consumer’s mobile device, as a contextual marketing element.
Spotting a Silo
Ask the data the right questions, the experts say, and it will provide answers. If a marketer has the suspicion that the best customers tend to research a product category via desktop, convert on a tablet, and pick up in-store, for example, it can use big data to find out for sure—and tailor campaigns, media buys, and inventory to that reality.
If the data can’t answer a question like this, there’s a more serious operational problem. “Any time you ask an analyst to look at X and Y, and the answer is that they can’t do that, the data is siloed,” Blickstein says. “When you are not able to creatively synthesize data and are limiting your ability to perform exploratory data analysis, your data is too siloed.”
The retailers surveyed by Monetate also rated the challenges they face in implementing their big-data goals. More than half (51 percent) said that a lack of shared data is an obstacle to measuring marketing ROI. “The telltale sign of siloed data is an inability to generate a complete picture of the enterprise,” Smith says.
Departments and personnel within a siloed organization have the tendency to withhold data or only ask about what they, specifically, need to know. “A good example is an organization that has their account services team or media managers come to the analysts just to do reporting for them,” Wisner says. “They look at the data as their own, not taking into account the wholistic view of the marketing funnel or loop.
Mobile has upped the ante. Marketers can now use location, provided by a consumer’s mobile device, as a contextual marketing element.
“Within an organization, there are siloes within functional areas that limit the use and allocation of data,” she adds. “If you have media buyers, for example, who are accustomed to doing media buys and planning the way they’ve always done them, [they] almost become proscriptive in terms of how to use data. Whereas if you are collaborative, you allow the analyst to come to you with all the data they can find, and expose all the trends that can be exposed.”
Leadership and Collaboration
To avoid silos, a person who’s ready to field information from every channel and use it to build ROI should lead the big-data strategy. “The effort should be led by someone who has a clear understanding of both marketing and technology,” Smith says. “It has to be someone who wants to listen closely to the market, can unify teams and systems, and ensure top-down support—someone who recognizes that all boats rise when access to information is flowing and accurate.”
New C-suite positions are emerging to lead the integration of the torrents of information coming in from every touchpoint. Seeing the need for a data champion, Home Run Media installed a chief data officer (CDO) to encourage the growth of a “culture” of data. “Once a culture of both data and refined curiosity is in place, a data-driven culture will emerge organically,” Blickstein says. “Creating easy accessibility to data and encouraging all divisions of the company to take the time to explore the data allows everyone to uncover their own insights.”
“A collaborative engagement between marketers and analysts is going to make them strong and grounded in data,” Wisner says. “There has to be [teamwork] between analytics, data scientists, and marketers to be effective. If you are solving for a business problem and want to come up with an answer, a meeting of the minds has to happen.
“Be brave and proactive in the way you handle your data,” she advises. “At the end of the day, if you are not moving the needle or showing growth in your business, you are not using data—big or small—in the right way.”