Actionable Analytics

October 1, 2013

Reexamining metrics forbetter data-driven decisions

By Michael J. McDermott

Brand marketers have a cornucopia of tools and channels that enable them to collect data with a level of accuracy and effectiveness undreamed of just a few years ago. But how effectively are marketers leveraging data analytics to make meaningful improvements to their marketing? The challenges revolve around what Patrick LaPointe, executive vice president of MarketShare, a Los Angeles–based global predictive analytics firm, refers to as the “human endeavor” factor as much, or more so, than technology.

Here’s an example of just how effective data analytics done right can be: One of MarketShare’s clients, a big consumer electronics firm, was measuring its advertising impact medium by medium, looking independently at TV, print, radio, and online to gauge each one’s impact on sales — an approach the majority of marketing organizations still use. According to Wes Nichols, MarketShare’s CEO and cofounder who wrote about this client in a Harvard Business Review article earlier this year, what was missing from that work was any analysis of how all these discrete channels interact with each other — such as whether a print or broadcast ad might prompt an online search that results in a clickthrough on a display ad and, eventually, a sale.

When the electronics company adopted a new data analytics approach to examine the interaction between channels, the results were eye-popping. YouTube ads, which accounted for 6 percent of the company’s ad budget, turned out to be twice as effective as TV spots, which accounted for 85 percent. And a full quarter of sales were being generated by search ads, which were getting just 4 percent of the ad budget. The company used those findings and the latest generation of predictive analytics to reallocate its ad spending, resulting in a 9 percent increase in sales with no increase at all in the total ad budget.

Today’s deluge of data can be a blessing and a curse. “The benefit is that we are getting closer to the holy grail of the 360-degree view of the customer and the ability to develop richer segmentation and predictive models,” says Alexander Edsel, director of the masters in marketing programs at the University of Texas at Dallas, who teaches graduate and undergraduate courses in digital marketing and advertising.

The ability to manipulate large volumes of data for analysis will allow marketers to develop real-time actionable models that integrate multichannel customer activities. “The challenges in getting there are many, however,” Edsel says. One of the biggest challenges is unrealistic expectations about how much work is required to create seamless solutions. “There will be some low-hanging fruit, but most actionable models need to be tested, refined, and validated, and that takes time,” he stresses.

Still, astute marketers are already finding creative ways to harness big data effectively to meet specific needs. Gilt Groupe Inc., for example, uses it to expand its relevance to consumers. “With the amount of data we have access to, there is no excuse for us not to provide a great, personalized experience,” says Tamara Gruzbarg, senior director of analytics and research at the online shopping site, which is headquartered in New York. With more than 8 million members at, the company has access to a plethora of data points about its customers — transactional information, of course, but also demographics, purchase patterns, brand preferences, site-visit timing and frequency, and which areas of the site they find most interesting.

Leveraging that voluminous cache of information requires a solid infrastructure. Gilt relies on a big, scalable data warehouse built and maintained internally so that its data engineering team can gather information from disparate sources (visitation, browsing, email) for relevant analysis. It uses a variety of tools, but continuous testing and optimization characterize everything it does. “We build the tools and test them in the market,” Gruzbarg explains. “We gather continuous customer feedback to ensure that what we do is really in our customers’ best interest.”

Gilt’s data analytics efforts focus on building a customer lifetime value metric for everyone in its database, which Gruzbarg says is “extremely useful in helping us understand where to direct our marketing efforts.” Her team tracks visitation and conversion patterns closely, paying special attention to channel preference. “Mobile has simply exploded as a shopping channel over the past year, and we are keen on understanding this better,” she says.

Significant advances in data collection and analysis technologies make muscled-up data strategies like Gilt’s possible, especially because of cloud computing and increasingly robust and reliable infrastructure platforms. “Analytics is getting more granular and more predictive simultaneously because technology can now be brought to bear at scale,” MarketShare’s LaPointe says. Moreover, it is transitioning from a world where it was basically done “by hand” and limited by how much iterative and explorative marketing one person could do. “Efforts had to focus on isolating just the most meaningful things, so analysis stayed at a high level, i.e., what is driving sales from paid media,” he says.

Putting Together Pieces of the Puzzle

Historically, data analytics has been backward-looking. Then “the marketing ecosystem started to fracture with the emergence of digital spend, social, and now mobile really coming alive,” LaPointe says. “The number of channels, touch points, devices, and ways of interacting are growing almost exponentially every 12 to 24 months.” He adds: “The old approach to data analytics — cranking it out by hand — simply was not going to work.”

As the models grew more complex, many marketers turned to outsourcing as a solution, but that approach has inherent limitations. One really good data analytics specialist in the U.S. can manage, at the most, three to five offshore professionals, LaPointe suggests. “Analytics is still primarily a human endeavor, and the models are only as good as the expertise and creativity of the individuals building them,” he says. “Once you start losing that direct management, your quality starts to dilute.”

While offshoring had limited success in advancing the effectiveness of analytics, cloud computing has been another story. The ability to distribute computational load across the cloud means marketers are no longer limited to building just one or two big models for a specific objective. “Now you can build hundreds and hundreds of little models that explain the relationship between A and B, reassemble them into a big picture of the world, and see if it is accurate and predictive,” LaPointe notes. And if that big picture does not resemble reality, marketers can cheaply and easily reassemble the data puzzle before them, through the cloud, until the right picture emerges.

Problem solved, right? If only. While data fluency has been improving, the resulting challenges are as significant as the opportunities. “CMOs are trying to make sense of disparate data sets and sometimes disconnected metrics coming from various silos of analytics, typically within different apps,” says Jeff Allen, director of product marketing at Adobe Analytics. “Down the org chart from the CMO, however, the data is not always making it into the hands of the individual contributors who rely on the data to optimize their work — arguably the most important consumers of analytics insights. The challenge and the opportunity is to drive a data-driven culture into the workflows of every marketer.”

However, there’s a stumbling block standing in the way of that goal. “The most significant  obstacle is the lack of education among current marketers to deal with these large data sets and, hence, the lack of resources to properly harness big data,” says William Rand, director of the Center for Complexity in Business and assistant professor of marketing and computer science at the University of Maryland’s Robert H. Smith School of Business. In response, some business schools have recently begun offering programs that specifically address data analytics in marketing, including one at Robert H. Smith.

The most important benefit of big data and its accompanying analytic tools is the power to build extremely robust and predictive models of how consumers behave, Rand believes. “With big data and the increasing trend toward interaction in the digital space, we can analyze everything a consumer does,” he says. “Not only can we understand what lift there is from a social media marketing campaign, we can even know exactly which users and what demographics responded to that campaign.”

Wilson Raj, global customer intelligence director at Cary, N.C.–based SAS Institute Inc., developers of analytics software, says data analytics in marketing can provide substantial benefits in four key areas:

  • Customer experience. Big data offers rich insight unachievable by examining customer feedback alone. “For instance, marketers can use operational data in call centers — such as wait times or time to resolution — to improve the customer experience across channels,” Raj says. “Operational data can also reveal training opportunities to enable front-line staff to deliver better service.”
  • Customer engagement. It is a daunting challenge for marketers to find out who their customers are, but those answers are essential. Data analytics can help marketers identify what their customers want and what marketers need to change.
  • Customer retention and loyalty. “Data analytics lets marketers augment existing customer touch points and anticipate new ones to help keep valuable customers loyal in a brand-fickle world,” Raj says. It can also help marketers allocate resources at the right junctures by developing more meaningful loyalty initiatives.
  • Marketing optimization and performance. As marketers shift budgets from traditional to digital channels, they need to be able to determine the optimal spend across multiple channels in order to be more accountable to the C-suite. “With a test-learn-optimize approach, marketers can deliver on the key determinant of business longevity: return on investment,” Raj says.

Big Breakthroughs in Data Analysis

“With the ability to measure, marketing can finally have an accounting value,” says Jennifer Ping, cofounder of Toronto-based Universal Insights Analytics. To translate marketing costs into associated revenue, marketers have to break down each marketing channel, look at past returns from past campaigns, and set clear goals for the ROI they hope to achieve from each channel going forward.

George Musi, head of cross-media analytics at DG, a global multiscreen advertising management and distribution platform based in New York City, acknowledges that the combination of information, stats, traffic patterns, and metrics needed to make better data-driven decisions will be different for individual marketers. However, all marketers want to know how their marketing activities perform across all channels. To get there, marketers need several data sets to develop the most appropriate and critical metrics for any marketing effort.

The scope of potential data-point sources is tremendous: Nielsen, Kantar Media, and comScore; post-conversion behavioral, awareness, and opinion measures; shopping and purchase intentions; CRM information; call centers; websites; transactional and point-of-sale data; search activity; social media; mobile engagement; clickstream; surveys; demographics; macroeconomic factors, and more. “Not every metric suits every situation or audience, and no measure or metric has meaning in isolation from a result or from other related metrics,” Musi says. “Additionally, it is critical that metrics be dynamic and evolve alongside business developments and changes.”

As analytics’ role has grown, chief marketing officers’ antennae have gone up, LaPointe says. Much effort has been chewed up in a search for the “magic” tweet or trying to determine the value of a like on Facebook, in large part because that is the kind of data most accessible to them. “It’s like the drunk looking for his keys under the lamppost. He didn’t lose them there, but that’s where he’s looking because that’s where the most light is,” he jokes. “Real insights come when you have a breakthrough in how to reach for numbers you don’t already have and drag them into the analysis — things like customer perceptions, competitive activity, changes in the buying process or channels or pricing.” Getting there, LaPointe notes, requires organizational adoption. “And that’s often the most difficult challenge to overcome,” he says. “The way to do it is through efforts based on credibility — by building credible analytics that will be accepted by key stakeholders internally, even those who might feel threatened by it.”


  • Analytics remains
  • a human endeavor, and the models are only as good as the expertise and creativity of the people building them.
  • Most actionable models need to be tested, refined, and validated.
  • Metrics should be dynamic and evolve alongside business developments.
  • It’s imperative to instill a data-driven culture into the workflow of every marketer.

A Case for Precision Marketing

By Ken Beaulieu

Sandra Zoratti, former vice president of marketing at Ricoh, believes it’s both an exciting and scary time to be a marketer. On one hand, marketers have a plethora of innovative tools to engage customers and prospects on all new levels; on the other, marketers have less control over messaging and branding in an age of customer power. Company transparency, she notes, happens whether it’s intended or not.

“Online conversations, not brand broadcasting, are the new determinants of a company’s brand identity,” says Zoratti, author of Precision Marketing: Maximizing Revenue Through Relevance, who was named Colorado’s Business Marketer of the Year in 2012. “Peer-to-peer conversations are increasingly powerful because we’re communicating with people, not business entities, and those same people make recommendations and shape brand perception. Leading marketers are learning how to engage and bond with customers and prospects by listening, asking, and sharing.”

Q. You define “precision marketing” as using customer-driven insights to deliver the right message to the right person via the right channel. How can marketing automation tools and platforms improve the relevancy of brand communications and increase response, revenue, and ROI?

A. Marketers have to speak in the language of business. Measuring ourselves in the context of revenue, response, and ROI allows us to show the concrete value of our investments and bridge our marketing activities to broader business goals. The need for marketing automation tools might not be obvious, but when you look at three unacknowledged, yet almost universal truths that the Chief Marketing Officer Council recently uncovered, the call to action becomes quite clear. The shared stumbling blocks to personalizing, targeting, and timing communications are insufficient data-gathering methods and analytics, siloed data that is rendered hard to obtain, and inaccurate and/or unclean data. This is where marketing automation has the potential to “save the day.”

Marketing automation tools and platforms lay the foundation for precision marketing, which is leveraging data-derived insights to improve marketing outcomes
by delivering the right message to the right person 
at the right time via the right channels. Marketing automation addresses the three common problems 
I just described by increasing the efficiency and accuracy of data gathering, analytics, targeted marketing message deployment, and measurement. To put it simply, marketing automation tools and platforms can help to identify the best outbound customer marketing by gathering inbound customer behaviors. They also help marketers overcome the “data paralysis” that holds us back from grabbing the gold ring of turning data into meaningful information that helps us to be relevant 
and engaging with customers — and pave the way for future revenue.

Q. How can marketing make a strong caseto procurement for implementing big data programs? 

A. Actually, this is the irony. Big data programs imply big dollars, lots of time, and infrastructure investments. With the plethora of cloud-based and modular solutions, there are great options for marketers who want to crawl-walk-run and implement, measure, learn, adjust, and repeat. That will enable the marketer to generate results that are relevant to the business, the customers, and the prospects and bring those facts to the CEO, CFO, CIO, and other executives. Since leveraging insights from data-driven techniques is a marathon, not a sprint, this is an excellent way to start and then expand.

Q. How can brands better bond with customers in an age of media proliferation?

A. Brands can strengthen bonds with customers in this tricky environment by conversing across all channels, being visible and transparent, listening socially, being responsive and relevant, and “dogfooding” — using their own tools and products to prove their value. The data gleaned from customer behavior provides the insights that guide this interaction, which is why it’s so important to use tools like marketing automation to divine the most meaning from the proverbial well of data that we’re standing on.

Ken Beaulieu is senior director of marketing and communications for the ANA.


"Actionable Analytics." Michael J. McDermott. ANA Magazine. Fall 2013.

You must be logged in to submit a comment.