Data and Measurement

As people conduct ever more of their lives as consumers and social animals online, they leave telling and durable traces of their behaviors, values, and inclinations behind them in digital repositories. Marketers have not been slow to recognize the opportunities presented by this information or to capitalize on them. The following selections offer:

•     Examples of some of the newest ways that brands are innovating their marketing approaches in response to these opportunities
•     Best practices for executing and measuring data-driven marketing that have quickly become standards in the industry

One particularly fruitful source of data for marketers has been the information regarding consumers’ locations and movements that mobile devices can yield. Office Depot, for instance, has leveraged this type of information to understand why outlets in certain places don’t succeed and to monitor which customer segments fail to find new locations when their nearest store folds. Brands can use location data to not only diagnose shortcomings but to seize opportunities as well. Panera Bread has demonstrated how such location data can be exploited in real time to deliver timely mobile ads to consumers when they’re in the vicinity of one of its restaurants.

The vast troves of personal data that are now available have enabled marketers to offer their messaging at the most opportune times and places and customize their offers to more precisely suit recipients. Nationwide, for instance, has analyzed customer data to tailor the insurance packages offered on its website to appeal to the specific attribute set represented by a given user.

The proliferation of mobile devices that has made location data so useful to marketers has also obsolesced older approaches to measuring the performance of digital marketing. The cookies that many marketers still use to evaluate the success of their campaigns were designed for a desktop world and don’t accurately track consumers’ use of mobile or behavior across multiple devices. This development illustrates the need for marketers to continuously and vigilantly respond to new technologies by updating not only their tactics, but also their approaches to measuring the performance of those tactics. As a response to the waning value of cookies, social media giant Facebook urged marketers to gauge the effectiveness of their marketing tactics by using conversion lift studies and split testing.

As the volume of data available to marketers has increased, so too have the regulations governing marketers’ collection, storage, and use of it. The EU’s enactment of the General Data Protection Regulation has subjected marketers to particularly stringent rules, whether they are based in Europe or not. This has put the onus on brands to ensure they are compliant with these and other applicable regulations.

As these examples indicate, the marketing discipline is continuously availing itself of digital data in extensive ways, presenting its practitioners with a growing wealth of both opportunities and challenges. This sea change, should not, however, persuade marketers that more traditional forms of acquiring consumer data are without value. For instance, Comcast continues to reap rewards from its customer survey panel while Procter & Gamble has used ethnographic research to inform a number of its successful product innovations. Thus, the future’s arrival foists the same dual mandates on marketers as it does on so many others, demanding that they have both the courage to fully capitalize on its benefits and the discernment to remain alive to the advantages inherited from the past.

The Data-Centric Organization 2018

“The State of Data 2017” (published in December 2017 by Winterberry Group, in partnership with the DMA and IAB’s Data Center of Excellence) revealed that U.S.- based marketers, publishers, and other data users invested $20.19BB on third-party audience data and related services and solutions in 2017.

The research shows how organizations are applying those investments to capitalize on the value inherent in data as a potential driver of decision-making, messaging, and superior customer experiences across all advertising and marketing functions.

Marketers and publishers remain intensely focused on achieving “data centricity” across their organizations, with 90.1 percent of survey respondents saying they either have or are planning to develop an enterprise strategy to support the use of audience data.

The same respondents said they expect those strategies will drive significant nearterm advances in how their organizations engage with data as an input to their advertising and marketing efforts. While only 9.8 percent of respondents described their organizations as “extremely” data-driven today, more than 44 percent said they expect to achieve that level of sophistication by 2019.

Amid that optimism, though, the same data users appear to be growing less confident that their efforts to date constitute real maturity given the growing complexity associated with deploying data. Fewer survey respondents described their organizations as at least “fairly data-centric” in 2017 (40.3 percent) than did the previous year (54.3 percent).

The most significant obstacle associated with achieving data centricity appears to be an industry-wide talent gap, particularly with respect to the analytics skills associated with data modeling, segmentation, and attribution.
Trending

Analytics, Organizational Structure, and Capitalizing on Location Data

With the help of Foursquare, Office Depot described some of the steps it has taken to develop its digital marketing capabilities and some of the benefits it has already reaped.

In the fourth quarter of 2017, Office Depot was in decline, closing stores and losing market share. In response, the brand began its transformation from a simple brick-and-mortar retailer of office supplies to a service-focused organization with more robust digital marketing capabilities than it had cultivated in the past.

To optimally execute these changes, Office Depot undertook two important organizational adjustments. The company:

•     Combined its marketing and e-commerce divisions under a single “E-Mark” umbrella
•     Consolidated ownership of and accountability for data and metrics with its Analytics Center of Excellence. This move increased the consistency of data, which in turn increased the speed of the organization’s decision-making

The efficiencies made possible by this reorganization, in conjunction with guidance provided by Foursquare, enabled the brand to capitalize on digital marketing opportunities that it had formerly ignored.

First, Office Depot reallocated its marketing budget so that two-thirds of the dollars that it once spent on print media such as circulars and direct mail were redirected to digital media. Second, Office Depot harnessed digital opportunities through the use of location data, which provided it with illuminating consumer reconnaissance. In one case, location data helped the brand discern that visits to one Office Depot location in the Great Lakes region had plummeted as a result of the closing of a neighboring mall that had helped attract shoppers to the area. Location data also helped the brand better understand which segments of the population it was failing with, as the places people go can serve as clues regarding their demographic identity. For instance, Office Depot’s analytics team analyzed the results of its own store closings to reveal that those brand loyalists who tracked down another location to shop at belonged to an older demographic, highlighting the brand’s weakness with younger consumers.

The use of location data, however, didn’t just help the brand diagnose challenges; it also helped it to seize opportunities. For example, in its effort to promote its copy and print services, the brand designed different creative materials to speak to different verticals, and, with insights drawn from location data, delivered them in places where their intended audience over-indexed. The effort yielded impressive results, producing:

•     A 200 percent increase in the brand’s click-through rate
•     A lift in store visits that exceeded the brand’s benchmark by 15 percent

Office Depot could tout these results as a vindication of the dramatic changes that it had made to its marketing approaches and its organization as a whole.

Up-to-the-Moment, Geo-Specific Mobile Marketing

Branded apps can be a particularly rich source of both location data and opportunities to press that data into the service of digital marketing. Visiting a website is a relatively anonymous way for consumers to contact a brand, and there is little expectation that their interaction with the brand will continue, much less evolve, after they close their browser tabs. The act of downloading an app, by contrast, is a more intimate way for consumers to contact a brand; it is, after all, a way of welcoming part of the brand to live on their devices. As such, it invites brands to reciprocate that familiar contact — for instance, by reaching out with individualized messages.

Thanks to their installation on personal devices, these mobile apps give brands access to troves of personal data, which, in turn, enables them to serve messages that are simultaneously and instantaneously customized based on users’ location and segmentation. Geospecific targeting enables brands to intercept consumers in mid-stride along their path to purchase.

Customer location isn’t the only newly significant data point for marketers; the specific moment when brands reach out to consumers is also critical. Dayparting no longer applies just to content consumed through radios or TV sets, as audiences increasingly rely on mobile devices. For a restaurant such as Panera Bread, the breakfast, lunch, and dinner customer are different from one another and should be segmented and appealed to accordingly.

Using Analytics and Personalized Solutions to Combat Default Bias

The default options offered to shoppers exert a powerful influence on their ultimate buying decisions, especially when the shopping experience takes place online, where consumers have fewer opportunities to ask questions that can guide their decisions.

Nationwide witnessed the effect that this phenomenon had on its own customers when it examined data on their approaches to buying auto insurance through its website. Traditionally, Nationwide’s website presented shoppers with Basic, Standard, and Premium packages as default buying options, which they could select as-is or customize.

Significantly, the Basic option appeared in the first tab that was visible to shoppers, giving it a prominence that made it a “decision anchor” — an option that influences choices even when it is not the one selected. The power of such a decision anchor was evident in the behavior of Nationwide’s customers: 25 percent selected the Basic option without adjusting it, but another 50 percent chose the Basic package and then modified it. Only 25 percent of customers selected the Standard or Premium options, whether in modified or unmodified forms, reflecting the deference commanded by the first option that customers see.

Default bias owes its power to a number of other factors besides the “decision anchor” phenomenon. These factors include:

•     Shoppers’ inattention: They may not even notice that other options exist besides the first one presented.
•     Aversion to switching costs: Once shoppers customize an option, they may be unwilling to invest the additional time and effort necessary to compare it with the customized version of another option.
•     Aversion to loss: Shoppers may feel that by selecting an option other than the default, they will lose quality or more money than they want to spend.
•     Deference to experts or normative choices: Shoppers may feel obliged to select the default because they assume that it reflects the wisdom of experts or the most common choice.

Though Nationwide could not revise human nature and its deference to default options, it found that it could use analytics and machine learning to personalize the options that it presented to online shoppers. Thereby, the organization was able nudge shoppers in the direction of choices that most adequately fulfilled their specific needs. This approach had the added benefit of decreasing the number of customers who downgraded from their previous insurance package when they bought a new one.

Using machine learning, Nationwide was able to identify the types of auto insurance purchases that were made by members of different sets of shopper attributes. Armed with this knowledge, Nationwide began presenting shoppers with a Basic, a Standard, and a Premium option that most closely reflected their attributes, which the organization gleaned from answers that they provided to questions on its website. Moreover, the website used the first visible tab to present shoppers with the option that members of their attribute-set were most likely to buy, whether that was Basic, Standard, or Premium.

In an experimental setting, Nationwide has since begun presenting customers with seven machine-learning-optimized choices instead of three to further personalize their shopping experience. The organization continues to evaluate the successfulness of this approach.

Best Practices

Protecting Data and Staying Out of Jail

To marketers, consumer data can seem like a substantial boon, but to those it’s extracted from, that data can seem like an equally substantial threat. According to a KPMG global survey, 55 percent of consumers said they had decided against buying something online due to privacy concerns. Furthermore, less than 10 percent of consumers felt they had control over the way that organizations handle and use their personal data.

A Pew Research study concurs with such data, finding that 76 percent of adults are “not too confident” or “not at all confident” that advertisers who record their online activity will keep those records private and secure.

Of course, there are regulations in place that attempt to address these concerns. In the U.S., the FTC aims to protect consumers against practices that are deceptive, unfair, or that unexpectedly reveal private consumer information, even if it doesn’t result in physical or financial harm. And it wages this defense of consumer privacy across all devices on which it might be threatened.

In passing the General Data Protection Regulation, the EU has adopted an even more strident posture, insisting that consumers have a fundamental right to access, erase, transfer, and restrict the processing of their personal data.

To help marketers whose data-collection practices might fall under the scope of these or other regulations, Reed Smith offered the following eight tips.

1. Be on the lookout for potential data liability issues. Risks can lurk in myriad areas of the law, including the Fair Housing Act, CAN-SPAM, employment laws, the Children’s Online Privacy Protection Act, and the Health Insurance Portability and Accountability Act. For instance, several housing authority corporations are suing Facebook for providing advertisers with a tool that allowed them to discriminate against people according to their housing. And the New York Attorney General settled a case with toy manufacturers who used a tool to track children online and target them with toy ads.

2. Protect any data that they or their vendors collect. To do this effectively, marketers must understand:

•     The quality of their vendors
•     The flow of data from beginning to end
•     Who stores data and where
•     Who has access and how
•     What steps are being taken to protect data

When marketers lack such an understanding, serious consequences can result. In one case a marketing firm inadvertently permitted public access to a database of 12,000 social media influencers, along with their home addresses, email addresses, and phone numbers, and the names of the brands that partnered with them.

3. Plan ahead for how to maintain access, control, and portability of their data. Maintaining access requires maintaining the ability to not only access data handled and/or controlled by third parties, but also the ability to discern:

•     Who is collecting data
•     The cost of the data
•     The quality, validity, and authenticity of data
•     How that data is stored

Maintaining control requires maintaining the ability to:

•     Manage how one’s data is collected and stored
•     Clearly document the rights to data
•     Discern limitations on data
•     Audit rights to data

For marketers to maintain the portability of their data, they must maintain the ability to:

•     Take data to another provider
•     Integrate their data with other data

4. Read and understand their own privacy policies. Liability for false advertising hinges on one’s privacy policies and other disclosures made to consumers. To avoid liability, marketers should examine how these disclosures characterize their practices for collecting, sharing, and using consumer data and consider any inconsistencies with their actual practices. For instance, Snapchat had to settle a case with the FTC after investigators discovered that it did not delete messages that users were told would permanently disappear.

5. Understand technology. When using data-collection technology, consider how it works, and if that would not just violate the law, but unnerve consumers. Unsettled web users filed several class action suits against Navastone after discovering that the organization was recording the keystrokes of its website’s visitors.

6. Use caution when collecting and using personal or sensitive information. TV manufacturer VIZIO was forced to pay the FTC more than $2 million after investigators discovered that it was collecting data on users’ viewing habits and selling it so other companies could target those users with tailored ads.

7. Participate in self-regulatory efforts and convince partners to do likewise. A number of groups organize such efforts, including:

•     Digital Advertising Alliance: A coalition of media and marketing associations and publishers and advertising service providers
•     Network Advertising Initiative: An association of advertising networks
•     Trustworthy Accountability Group: A joint marketing-media industry program, which focuses on four initiatives: eliminating fraudulent digital advertising traffic, combating malware, fighting ad-supported internet piracy to promote brand integrity, and promoting brand safety through greater transparency

8. Partner with IT, data security, and legal. Owing to their position on the cutting edge of the technologies used to reach and monitor consumers, marketers are in the best position to flag potential issues for legal and data privacy specialists.

Improving Digital Metrics: Recommendations from Facebook

Outdated Measurement Technologies: Cookies and Clicks

Cookies were designed for a desktop world, and don’t accurately work on mobile or across multiple devices. Facebook has found that, as a result of these shortcomings, cookies can:

•     Overestimate reach by 58 percent
•     Understate frequency by 135 percent
•     Miss broad age and gender targets 46 percent of the time
•     Miss up to 66 percent of conversion events

It was in a similarly pre-mobile time when marketers seized on clicks as a valuable metric — a time when consumers searched, discovered, and purchased on a single device. Today, consumers proceed through these stages of their customer journeys across an array of devices and channels. Consequently, a click on one device may, undetected, result in conversion on another device.

Measurement Solutions

Facebook recommended conversion lift studies as one method that marketers can use to assess the effectiveness of an online ad more accurately. In this model, advertisers identify a business objective and an audience that they want to reach. The advertiser then takes a randomized sample of that audience and splits it into a test group and a non-overlapping control group. The test group receives the ad while the control group does not. In partnership with the media publisher or a platform such as Facebook, the advertiser compares conversions in the test and control groups to ascertain the ad’s effectiveness.

Marketers may also want to measure the effectiveness of a particular feature of an ad, in which case Facebook recommended using split testing. In this model, the advertiser assembles a test and a control group, giving each group an ad that is identical to the one seen by the other group except for one detail or “variable.” For instance, the ad seen by the test group might be narrated by another actor. The difference in the two ads’ conversion rates will quantify the value of the variable.

Case Study

AVEENO Capitalizes on Loyalty Card Data

Marketing Situation

AVEENO’s largest product segments were generating increasing revenue, but the brand’s lesser-known product categories, such as anti-aging and hair care, were experiencing declines. AVEENO set out to capitalize on the success of its heritage product lines to support those lesser-known lines by driving customers to purchase its products across multiple skin care categories.

Insights

AVEENO found that its loyal buyers were making purchases from a skin care company in five of the categories that it sells in; however, they were only purchasing AVEENO products in 1.4 of those categories. The brand further discovered that 81 percent of its buyers were only purchasing in one segment. This imbalance presented the brand with an opportunity. The AVEENO baby buyers were a particularly favorable target, as they tended to spend more than the average buyer, and two-thirds of them did not yet purchase products from the AVEENO adult categories. The brand determined that to generate continued growth, it needed to drive retention and loyalty with its core consumers and ensure that all AVEENO products were top-of-mind for them.

Key Tactics

Mining loyalty card data, AVEENO and Neilson Catalina segmented the brand’s established customers into audiences that did not purchase from particular AVEENO product lines: body wash, body lotion, facial cleanser, facial moisturizer, and anti-aging skin care.

Once the audiences were defined, AVEENO customized its messaging based not only on the audience segment, but also on real-world triggers, such as the UV index, retailer locations, the season, and other contextual details. The result was more than 18,000 pieces of unique and tailored creative that reached consumers at the times they were most ready to buy.

AVEENO’s marketers were guided by their experience on previous campaigns. The brand knew from its 2016 performance that its audience read material from premium content outlets, prompting it to test new media-buying methodologies to gain more presence on sites that performed well in its previous campaign. AVEENO vetted and chose publishers based on both scale and relevance and executed its customized cross-category strategy with premium sites to scale its audience segments and identify users while they were on high-quality content.

The brand partnered past experience with emergent technology to maximize the campaign’s success. With the assistance of Ansa RSi (an automated data analytics service), AVEENO optimized its targeting in real time based on SKU-level sales. Ansa RSi then supplied additional reporting that revealed stores with increased sales lift to enable further fine-tuning of the campaign.

Campaign Success

AVEENO’s 2017 campaign resulted in a $12.06 Return on Advertising Spending (ROAS), exceeding benchmarks by 359 percent and resulting in total incremental retail sales of $4.8 million — a 5 percent increase. Moreover, the ROAS in 2016 and 2017 remained the same despite lower spending per household due to a 27 percent increase in reach.

Comcast Hits Mark with Customer Survey Panel

Though big data continues to prove itself an important source of marketing insights, it cannot, according to Comcast, replace the value of customers’ answers to direct questions. To most effectively gather this kind of information, Comcast has spent the past six years developing a resource now called the “Xfinity Insights Community.” This survey panel consists of more than 150,000 current and former Comcast customers who provide the organization with real-time answers to core business questions, which help guide the brand’s strategic decision-making.

Comcast populates this survey panel using email campaigns and the offer of incentives such as gift cards for participating. When respondents join, they submit profile questionnaires that supply Comcast with information, including their age, income, education, ethnicity, Comcast product/service type, and TV-viewing behavior. Just as importantly, Comcast, with participants’ permission, maintains access to set-top box data, including its records of viewing history, for 80 percent of the Xfinity Insights Community.

Armed with this information, Comcast’s surveys match each respondent (with their address and account number) to the full history of their customer behavior, thereby assuring the organization of the integrity of its data. Maintaining its own customer survey panel provides Comcast with several other advantages as well, including:

•     Speed: Comcast can compile survey results in a week’s time.
•     Granularity: The sheer size of Comcast’s survey panel enables it to identify and target narrow segments.
•     Cost savings: Generally, Comcast only has to pay for the maintenance of its panel, not the more expensive external sample.
•     Marketing opportunities: Panel activities give Comcast the chance to not only question but also market to participants.

Though these benefits are valuable, they do not preclude Comcast from using other resources. As the need arises, the organization will also make use of external panels — for instance, when it would like blinded respondents to answer questions on, say, price-sensitive issues.

The Xfinity Insights Community proved especially useful to the brand as it worked to piece together competitive intelligence on alternatives to cable television such as subscription video on demand (e.g., Netflix and Hulu) and live linear TV (e.g., Sling Television). The responses of panelists helped Comcast determine how large the demands for these services were and in what regions they were the largest. In addition, the survey data helped diagnose the causes of “cord-cutting,” revealing that the behavior, which several years ago was confined to tech-savvy millennials, is now being adopted by people from a wide variety of demographics.

P&G’s Approach to Ethnographic Research

Whereas Comcast identified important benefits in the use of survey panels, Procter & Gamble touted the value of ethnographic observation. This preference reflected Procter & Gamble’s belief in the inadequacy of simply asking consumers about their behaviors, as behavioral scientists have demonstrated that survey respondents often supply misrepresentative answers. To work around this tendency, brands must directly observe and listen to consumers in lived contexts to acquire reliable insights.

To illustrate its approaches to capitalizing on such data, the brand offered two examples. In the course of conducting ethnographic research on its female consumers, Procter & Gamble observed that these women do not leave their deodorant in the bathroom; the product accompanies them throughout their day in their cars and in their handbags. Nonetheless, these women can feel inconvenienced or self-conscious about carrying around a large stick of deodorant, so Procter & Gamble responded by developing Secret Freshies, a golf-ballsized deodorant dispenser that can fit comfortably in even the most petite clutch purse.

Similar ethnographic research has influenced product development in Procter & Gamble’s laundry division. Researchers observed that female Japanese consumers vigorously shook their clothes after taking them out of the dryer to minimize wrinkles, but these women worried that this would damage the fabrics. In response, the brand developed a wrinkle-reducing laundry pod.

Source Information

“Analytics, Organizational Structure, and Capitalizing on Location Data.” Gayle Fuguitt of Foursquare; Darren Zap, senior program manager at Office Depot; Ashley Freeman, senior program manager II at Office Depot. ANA Data and Measurement Conference, 9/12/18.

“Up-to-the-Moment, Geo-Specific Mobile Marketing.” Carol Wolowic, senior manager of media at Panera Bread; Gayle Fuguitt, chief of customer insight and innovation at Foursquare. 2018 ANA Media Conference, 2/28/18.

“Using Analytics and Personalized Solutions to Combat Default Bias.” Bob Forrest, director of data dcience at Nationwide. Data & Analytics Members Only Conference, 5/30/18.

“Improving Digital Metrics: Recommendations From Facebook.” Dominic Williamson, marketing science lead at Facebook. Data & Analytics Members-Only Conference, 5/30/18.

“How to Effectively Measure Conversions from Multi-Touch Campaigns.” Roger Ares, SVP of marketing analytics and client insights at SunTrust Bank. ANA Data & Measurement Conference, 9/14/18.

“Protecting Your Data and Staying Out of Jail.” Keri Bruce, partner at Reed Smith LLP. 2018 ANA Advertising Financial Management Conference, 5/1/18.

“Better Together.” 2018 Internationalist Innovation in Media Awards Grand Prix Winner. Brand: AVEENO. Lead Agency: Caldreon; UM.

“Building, Managing, and Capitalizing on a Customer Insights Community.” Ginny Too, executive director of marketing, strategy, and planning at Comcast. ANA Data & Measurement Committee Meeting, 3/15/18.

“Understanding Consumers’ Needs Through Data: Procter & Gamble’s Approach.” Kirti Singh, chief analytics and insights officer at Procter & Gamble. ANA Data and Measurement Conference, 9/13/18.

“The Data-Centric Organization 2018.” DMA, 2018.
Source

“Data and Measurement.” Insight brief compiled by Morgan Strawn, Manager of Content Creation, Marketing Knowledge Center. Designer: Amy Zeng, Marketing and Communications, ANA. Editor: Ken Beaulieu, Vice President of Marketing and Communications, ANA. © Copyright 2019 by the Association of National Advertisers, Inc. All rights reserved.