Online Exclusive

Smart Mail Uses Big Data

Optimizing content relevancy by using business intelligence and information management, also known as “Big Data,” can be part of a powerful, value-added proposition for print service providers and their customers. Market research firm IRI (Information Resources, Inc.) has been behind grocery-store barcode scanning technology for decades in the consumer packaged goods (CPG) space. In its more basic form, predictive analysis is what Netflix uses to recommend movie genres to its monthly subscribers: “People who rented this DVD also liked …”

When helping client to market products during this holiday season, make sure they’re “not only generating traffic but … capturing customer data in efficient and meaningful ways,” advised custom landing page and analytics software platform Dukky in a late October blog post, “so when the holidays are over you’ve seen growth and an influx that can impact your marketing all year long. More people, more Facebook likes, more followers, more engagement, more influencers, more research data -- all of these things add up, giving you a better platform for next year’s marketing.” Dukky was part of Adweek’s media buy of the year and PODi’s Best Practices for harnessing the power of social media.

Low-level analytics may involve between 1,000 and 5,000 patterns, explained John Leininger, professor of graphic communications at Clemson University in South Carolina. Clemson’s College of Business and Behavioral Science offers a master’s degree in marketing. Other prestigious universities, including Northwestern, offer master of science in predictive analytics that delve much, much deeper. To more effectively market to consumers, almost every major retailer -- from grocery chains to investment banks to the US Postal Service -- has a “predictive analytics” department devoted to understanding not only shopping habits but also personal habits.

Target Corp. statistician Andrew Pole has a master’s degree in statistics and another in economics. “Target has always been one of the smartest at this,” Eric Siegel, a consultant and chairman of a conference called Predictive Analytics World, told the New York Times. “We’re living through a golden age of behavioral research. It’s amazing how much we can figure out about how people think now.” Sometimes, how the collected data is used can prove too powerful, as most attendees at the annual DMA (Direct Marketing Association) show in Chicago last month know. Target may have taken its data too far two years ago, when one of its stores figured out that a teen girl was pregnant before her family knew. The goal was to sell pregnant women baby items they did not even know they needed yet. With an accuracy rate of approximately 80 percent, Target is able to predict that women in their third trimester of pregnancy will purchase certain products, such as cotton balls and unscented body lotions. “We are very conservative about compliance with all privacy laws,” Pole told a Forbes reporter at the time. “But even if you’re following the law, you can do things where people get queasy.”

Nonetheless, predictive analytics, data blending, real-time queries, and free-form search are top priorities for active big-data deployments in 2014. Implementations in production rose to 34.3 percent this year from 27 percent in 2012, according to “Big Data: Operationalizing the Buzz,” a study released in late October by Pentaho Corp., Enterprise Management Associates (EMA), and 9sight Consulting. In addition, 68 percent of companies are running two or more projects as part of their big-data initiatives. Taking advantage of batch processing in open-source software frameworks, such as Apache’s Hadoop, can help businesses make more accurate predictions, leveraging big data for fraud detection, upselling, and a range of predictive applications, the study said.  “In 2013 we have seen customers and prospects moving beyond single Hadoop deployments to hybrid data systems that might include Hadoop for scalable storage and MongoDB for high speed front-end analysis, STORM for real-time analytics, and an engine like Cloudera Search or Apache Search for quick data look up,” noted Richard Daley, chief strategy officer at Pentaho.

EMA big-data analyst John Myers added, “We are entering the era of hybrid data ecosystems where relational, structured data must be able to interact with non-relational data. Organizations have multiple data platforms and more sophisticated users,” Myers continued. “As they begin to mash up information in these new multi-structured data platforms, ecosystems management, data access and integration, and data analytics become much more important.”

Analyzing the numbers

As Clemson’s Leininger told, “There is a big difference between being a number cruncher and actually understanding what you are getting out of those numbers.” That’s where the analysis comes into play. Marketing with consumer data really is not a new strategy, after all. Citing the book Strategic Database Marketing, by Arthur Middleton Hughes, as the “Bible” of the industry, Leininger added that this type of thinking has been around since the 1940s, “long before digital presses.”

He pointed out that last month, some 125 teachers gathered at a pre-DMA meeting of the non-profit Marketing EDGE foundation, formerly the Direct Marketing Educational Foundation (DMEF). On the agenda was Hughes and Jim Sellers’ Recency, Frequency, Monetary Value (RFM) analytics. The value of RFM analysis as a method to identify high-response customers in marketing promotions -- and to improve overall response rates -- is well known. One common example is that women typically cut their hair every four weeks; hence, the best time to send out coupons is at the end of the third week, so they can schedule accordingly. Less widely understood, however, is the value of applying RFM scoring to a customer database and measuring how customers migrate from over time.

Lists can be purchased for less-sophisticated efforts, said Leininger. “There’s a printer in Greenville who did product scoring with Latinos,” the professor cited. “They bought Census data and are using the Every Day Direct Mail program” from the USPS. Similar list services can be bought from firms such as Infogroup and USAData. As sales increase, “you can then bring the [analysis] process in-house for higher revenues,” Leininger told an audience during his 90-minute seminar, at PRINT 13 in September, entitled Predictive Analysis: Helping Your Customers Target Their Clients. “Where do you start, and who is the right person in your company to make this work? Do they even work in your company right now?” he asked in his introduction, providing the answers later in the session.

“Do you know which of your customers is likely to make a purchase of services from you this month?” he continued. “The best way to know that is to know what … customers are going to do. A common model used in the marketing industry deals with ‘Recency, Frequency, and Monetary Value’ [the aforementioned RFM], but this is only one option and strategy. It is all about data management and data analysis.”

Revenue-generating analytics

“These days, businesses have access to more customer data than ever before, but marketers often struggle to capitalize on this information in a timely or accurate manner,” Robert Youngjohns, senior VP and GM of HP’s Autonomy product, blogged this past October, when the Digital Marketing Hub was introduced. The hub is a new solution that brings integrated and advanced customer analytics to marketers.

“[It] is a great example of HP’s breadth and depth, and our ability to bring together the best of HP to build comprehensive technology solutions for the new style of IT [information technology],” Youngjohns continued. “The HP Digital Marketing Hub is also a major milestone in HP’s HAVEn big-data strategy, leveraging HP Autonomy’s IDOL and HP Vertica, groundbreaking algorithms from HP Labs, and the HP Converged Cloud, to process massive volumes and varieties of customer data at scale and in real-time.” (See sidebar.)

Analytics are an effective tool to get a “birds-eye” view of what consumers want, but they also can be used to zoom in on specific groups of customers, allowing for the more effective segmentation and targeting of promotions. For example, home improvement chain Stine Lumber in Louisiana used Dukky, a promotion builder and tracker, to create a multichannel Father’s Day campaign to grow its customer database. Data collected on gender, age range, and channel preferences enabled the company to create marketing strategies targeted to a younger age group, ensuring a new generation of loyal Stine Mart customers.

Its Ultimate Father’s Day Giveaway offered the chance to win a new grill and a collection of other tools and gifts for dad. Stine’s marketing strategy included television commercials, radio, in-store advertisements, direct mail, email, and local blog communications. All respondents were driven to Stine’s Dukky promotion as the portal to capture customer data, engage with new and current customers by asking three simple questions, and incentivizing participants to share the offer with friends via email and more than 340 social networks. Results doubled over expected results with direct marketing alone, as the Dukky online promotion received more than 28,000 visitors from all media channels, with almost half of all traffic entering the campaign’s landing page through links shared on social media networks. Facebook and Twitter were the most effective sharing outlets, with share-to-response ratios of 1:4 and 1:10, respectively.

Target-market analytics revealed that nearly half of those who entered the giveaway were between the ages of 41 and 60. Specifically, men between the ages of 51 and 60 entered the sweepstakes more than any other demographic. Stine used the poll feature to learn how to communicate more effectively with current and new customers, particularly on which product categories they would be most interested in receiving emails about for special deals and promotions. Twenty-five percent of customers were interested in receiving lawn/garden offers, while 15 percent wanted to receive deals on tools and 13 percent on home decor. This data is being used to target customers and offer them more relevant promotions and products going forward.

Data captured in analytics also makes it easier to keep courting customers who do not redeem the offer the first time. For example, Dukky client HOM Furniture held a special sales event that lasted an entire weekend. Customers who did not redeem during the event weekend were contacted to set up future appointments. The analytics showed the sales team which department the customer was considering purchasing a product from, so the appropriate sales person was able to make the call.

Related Reading:

Direct Mail: The Good, the Bad, and the Scary

HP White Paper (PDF): “Put Big Data to Work Increasing Conversions and Marketing ROI

Data Beyond Print

As customers interact over an increasing number of channels, the need for multichannel analytics to gain insights from Big Data has become the top priority for chief marketing officers, said Hewlett-Packard. Marketers are also challenged by the ineffectiveness of products that are either too limited in scope, require significant amounts of time, or consume too many resources. HP Digital Marketing Hub enables customers to turn Big Data into insight-driven action by removing many of the challenges associated with advanced multichannel analytics. The product helps to create personalized experiences for customers based on deep insight into their cross-channel behavior. Product features include:

  • Get a complete picture of the customer: Combine transactional data such as customer relationship marketing (CRM), purchase history, web analytics, and email metrics with human-friendly data such as call logs, emails, social posts, and comments to get a holistic view of the customer using an intuitive dashboard.
  • Get real-time customer insight: Use advanced analytics to identify key customer segments and attributes for specific key performance indicators (KPIs), all delivered in real time and without the need for technical expertise.
  • Leverage the HP HAVEn Big Data platform: Get superior analysis and insights on a massive scale combining any data source, structured or structured, in real time and in a secure environment.
  • Understand the entire customer journey: Turn disparate data from different channels into a single view of the customer, mapping relationships between channels to understand their influence on the customer journey.
  • Create multichannel personalized experiences: Instantly identify offer positioning and messaging that resonates for specific segments, and increase conversions by automatically creating multichannel experiences using similar content.
  • Improve your channel strategy: Apply advanced algorithms specifically geared towards marketing problems to optimize your channel strategy so you create the best customer journey for selected KPIs.

Data for digital

PointRoll is another company (owned by Gannett) that helps advertisers, agencies, and publishers to create, deploy, measure, and optimize interactive and action-inspiring digital campaigns across channels and devices. OnPoint is its new campaign management, analytics and ad delivery platform -- a singular tool that lets marketers reach the right audience with relevant, consistent, and creative advertising across multiple digital channels, while easily monitoring and optimizing campaign performance. OnPoint combines the ability to create more engaging campaigns with unmatched formats and features, dynamic creative optimization, and ad analytics to provide a consistent, relevant consumer experience across in-stream video, display, rich media, mobile, social, tablet campaigns, and more.