Marketing Data Fundamentals: What CMOs Need to Know

. August 15, 2017
In the video above, marketers from IBM, NASCAR, Nationwide and Hyundai discuss data challenges.

Data is at the core of almost every key challenge that CMOs — or any marketers for that matter — face today, and will likely have a profound impact on marketers’ ability to prove their value.

“CMOs must understand how to find and reach their target customers, which is a data problem,” says Michael Schoen, general manager and VP of marketing solutions at Neustar, a global information services provider . They must also understand, he adds, how effective their efforts are in media spend, message optimization, and channel optimization, as well as the types of customers who engage with their brands and buy their products — all data problems.

One data aspect that’s sometimes overlooked, Schoen says, is the ability it gives CMOs to forecast the brand’s future and lead it forward. “It’s about understanding, through data, how the market and consumer behavior are changing, and how those changes need to be reflected in product development,” he adds.

So how are CMOs doing when it comes to mastering data? In general, not well it would appear.

Schoen gives them a collective C-minus, although he’s quick to add that some are doing much better. The low grade isn’t because CMOs are ignorant of the problem. “In part, it’s because the opportunity data presents to CMOs is so vast, and the capability of technology has grown faster than even the best organization’s ability to keep up with it,” says Schoen.

A recent CMO Council study highlighted the scope of the problem. Of particular note, 78% of marketers surveyed agreed that the CMO should be the catalyst and driver of a brand’s data-driven customer strategy, but only 19% said they currently are. This prompted CMO Council Executive Director Donovan Neale-May to warn in a subsequent interview that if CMOs don’t get better at dealing with data, their jobs could be in jeopardy.

‘Without Insights, It’s Only Information’

Data, in a broad sense, is nothing more than facts and statistics collected for reference and/or analysis. The main characteristic of data is that it represents objective reflections of the real world, as opposed to opinions and beliefs.

Nevertheless, data can be incorrect, biased, or incomplete, says Dirk Beyer, head of data science research at Neustar. “The goal is to use data for objective and informed decision making and communication, and to establish accountability,” he says.

In today’s digital world, almost everything a business or customer touches produces data. Companies have access to an avalanche of data about transactions, customer behavior, supply chain activity, sales performance, product use, and more, says Rishi Dave, CMO at Dun & Bradstreet. “But that data is not relevant without insights; it’s only information,” he says. “Data needs to be captured, consolidated, and normalized, and then it can be analyzed for insights.”

Data can be categorized many different ways. Cory Treffiletti, VP of marketing and partner

Demographic data is one of four basic data buckets.

solutions at Oracle Data Cloud, recommends putting data into one of four basic buckets:

  • Demographic data. Historically the most common data type used for targeting, often with broad generalizations.
  • Behavioral data. Highlights interests, intent, and consumption. Treffiletti breaks this down into two subcategories: behavior that indicates broader interest in a topic (e.g., reading articles) and in-market data derived from behaviors farther down the purchase funnel (such as comparing features and prices).
  • Geographic data. Pinpoints a user’s physical location. Common sources of this type of data include offline, IP, latitude/longitude geographic tools, and beacons.
  • Purchase-based data. Derived from actual purchases.

But as Beyer says, it’s difficult to create a complete taxonomy of all the data a brand might want to use to run its business. He cites these broad categories as particularly relevant:

  • Sales transactions data. “This is the ultimate outcome of the business,” Beyer says. “If the business has a direct sales relationship with the customer, this data is very granular and can include all aspects of the individual transaction.”
  • CRM data. Encompasses all the data that describes a particular customer, including historical sales, data the customer has shared with the marketer (surveys, subscriptions), customer service history, and any relevant data from other sources.
  • Marketing execution data. Includes records of marketing interactions in each channel used, as well as earned and owned interactions with customers. This data can be at the individual customer-touch level, or it can be an aggregated level of estimated impressions in a given market. In the latter case, cost data and customer response data are important elements.
  • Market conditions data. Describes the economic environment, competitors’ activity, weather conditions — anything related to the background of the marketer’s own business activity. “Marketers might not be able to influence these things, but they can help explain the business outcomes,” Beyer says.
  • Supply chain data. Encompasses critical information about the organizations, people, activities, and resources involved in moving a brand’s products and services from their source of origination to the end customer. It’s important because marketing must execute within the constraints of what can be produced and delivered. Marketing also can provide valuable feedback to product design or production and to shipment plans.
  • Customer sentiment data. Often obtained through surveys and focus groups, this data helps marketers understand customer attitudes toward their products, brands, and advertising messages.

Building a Data Infrastructure

Regardless of the type of data, or the purpose for which it is being used, the effectiveness of data-based marketing initiatives is highly dependent on the quality of the data itself.

“The first thing that is required to use data effectively is to have a data strategy,” Schoen says. An easy mistake to make is to develop a strategy based on low-quality data. “This is definitely one of those GIGO [garbage in, garbage out] areas,” he says. “You may have the best process, strategy, and algorithms for your data-based decision making, but if the data is wrong, you’ll end up doing the wrong thing, making wrong decisions, targeting the wrong consumers.”

There are many ways to define data quality. D&B starts by asking four questions:

  • Is it accurate?
  • Is it complete?
  • Is it timely?
  • Is it globally consistent?

“If you can answer yes to all of these questions, you likely have taken the steps to ensure your data quality,” Dave says. He adds that it’s imperative for marketers to have a process in place to maintain quality because “data becomes outdated literally within minutes. If you aren’t constantly cleansing your data and appending it with trusted third-party data, the quality will rapidly decline.”

Data quality is Oracle Data Cloud’s primary concern because it is fundamental to the success of its customers’ and partners’ businesses, Treffiletti says. Its data quality assurance process focuses on six attributes:

  • Seek out qualified data that has been verified across multiple sources.
  • Look for large data sets from your choice of data providers.
  • Consumer data across all connected devices provides the most holistic view.
  • Ability to activate a robust array of IDs, including cookies, registration ID, and device ID assets.
  • Ensures the company has access to a large pool of media partners to activate a data strategy.
  • Does this data type make sense for your marketing initiative? Does it align with your KPIs and target audience?

Sources of Data

Another important filter for parsing data quality is source of origin — whether it is first-, second-, or third-party.

First-party data is information that marketers obtain directly from customers through their interaction with the marketer’s brand. Loyalty card programs are a common source of this kind of data, and it’s generally the most valuable. It includes personally identifiable information (PII) such as name, address, phone number, email address, and, often, an extensive transaction history.

Second-party data typically comes from publishers. “At Neustar, we think of second-party as any intermediary between the brand and the consumer,” Schoen says. For example, if Starbucks were launching a sports drink, it might be interested in data on users who view, say, track and field content. The data is relevant and presumably accurate, but it may lack PII and is not owned by the marketer.

Third-party data comes from an unrelated vendor, is invariably associated with anonymous digital identifiers, such as cookies, and rarely includes PII. It’s most often used for building and understanding marketing campaign audiences. Third-party data providers often use modeling to replicate small-audience attributes in larger samples, which can dilute accuracy.

Marketer confidence is generally highest in first-party data (especially of the sort generated by loyalty card programs) and lowest in third-party data.

Second-party data can vary considerably. Schoen offers an example: “If Starbucks were doing a marketing partnership with Safeway to target consumers who purchase coffee in Safeway stores, it would likely have a high level of confidence. It’s still second-party data, owned by Safeway, but Starbucks knows that Safeway has high-quality data about what its consumers are buying. They might have less confidence doing the ESPN deal because ESPN doesn’t have the same kind of PII on its users.”

Identity As Key Metric

From a marketing perspective, the single most important data attribute may well be identity. Records of customer interactions often contain PII as well as semi-anonymous identifiers such as cookies and Ad-IDs.

“Being able to tie these identifiers together, so that a more complete picture of the customer and the marketer’s interactions with him or her can emerge, establishes an identity,” Beyer says. Identity is critical because it’s the way that everything gets connected, and it’s often a way marketers determine data quality.

One of marketing’s biggest challenges today is not only knowing and understanding customers across each of their individual devices (the average person uses five devices and has more than 25 IDs), but connecting them to one consistent message. An effective customer graph connects all those identities to form a single view of an individual. This allows marketers to seamlessly reach their audiences across multiple channels and accurately measure campaign performance.

Connecting the Dots

But data’s greatest value to marketers may be its ability to improve customer experience. “Customer experience is the ultimate strategic differentiator in the foreseeable future,” says Wilson Raj, global director of customer intelligence at SAS. “The only way you can differentiate is by delivering a unique customer experience that’s based on a solid, connected business strategy driven by marketing analytics.”

To maximize data’s value, marketers should view customer experience as one strategy across all media. “Keep it focused on the customer and the context in which the customer operates,” Raj says. “Use your data to maintain that focus, continue to enrich your existing data with new sources, and align the marketing process with the customer journey.”

Best Practices for Data Management

Here are a few data best practices from the experts interviewed for this article:

  • Have a well-defined customer data strategy. Determine the problems that need solving, such as customer acquisition, improved customer lifetime value, reduced churn, and increased engagement.
  • Grow data capabilities incrementally. Start with projects where success can be documented with clear metrics, and build on wins.
  • Eliminate data silos. Make sure all departments across the organization are using the same networks and software solutions, or at least ones that are completely compatible with each other. Relevant data should be accessible to any department that can use it, not considered the “property” of the originating department.
  • Set data-based analytic triggers to ensure brand activity and contextual relevance throughout the customer journey.
  • Align data to campaign KPIs.

Michael J. McDermott writes about entrepreneurship, finance, technology, and — especially — marketing for both print and digital publications.

A version of this article first appeared in the ANA Newsstand. Published here with permission of ANA.


Category: Articles, CMO Briefings, Data, How-To

About the Author ()

Leave a Reply

Your email address will not be published. Required fields are marked *