Creative & Tech

360i Report: The CMO’s Guide to Big Data

November 25, 2012

Executive Summary
It is rare that Hollywood makes a movie about analytics. It is rarer still to find business managers who look like Brad Pitt. “Moneyball” is a movie about the successful application of analytics in major league baseball by Oakland A’s General Manager Billy Beane (Pitt) in the early 2000s. He built a winning club on a relatively small budget by leveraging information his competitors ignored. This report explains how the “Big Data” explosion currently underway in corporate enterprises can create similar opportunities for marketers to outflank competition and gain advantage in digital marketing. Alternatively, marketers who are slow to capitalize on the signals that their customers and the market are sending them — potentially buried in Big Data — are at risk of being outmaneuvered.

In this report
What is Big Data? Big data refers to the explosion in information coming from diverse sources–in particular, machine-generated data from mobile devices and data from social media–that is too large to cost-effectively manage and analyze using traditional IT techniques.
Why does it matter? Marketers who successfully measure what matters within this influx of data will realize greater marketing efficiency and ROI, and identify opportunities to advance their brands’ offerings and positioning.
What are the associated challenges? With so much information available, figuring out what data matters and how to harness it is the first and greatest challenge. In addition, CMOs must rally their teams to break down silos, share data and work together to maximize use of it.
How can I get started? Appoint a Chief Data Scientist or Chief Marketing Scientist and develop a team. Then, create an internal plan to educate the marketing team on your data — what’s available and how to use it. Finally, internal capabilities and technologies are needed to help you take full advantage of Big Data.

Defining Big Data
In its simplest form, Big Data can be defined as large volumes of information, including both structured and unstructured data. Structured data is found in traditional enterprise databases or data warehouses. Unstructured data includes raw machine-generated data, data from personal productivity applications such as email, word processing, and presentation software, and rich mixed-media data from social networks.

Big Data is the result of two recent trends:
Proliferation of data in both size and type. The growing popularity of social media and mobile devices has resulted in vast new archives of customer data. Consumers spend almost $300,000 a minute shopping online; brands receive 350,000 “likes” per minute on Facebook; and Twitter users send more than 600,000 tweets per hour. Six hundred million more people in the world own mobile phones (4.8 billion) than own toothbrushes (4.2 billion). Mobile devices with GPS capabilities offer “location” information that can be very useful to marketers. And social media provides rich archives of data about consumer behaviors and attitudes towards brands that is invaluable.
New technologies specifically designed to manage and analyze Big Data, such as Apache Hadoop and Apache Hadoop MapReduce, have helped companies more fully leverage information in real-time.

Add all of these new sources of information to the petabytes of data already being generated within the ecosystem of a typical Fortune 1000 company, and you can find yourself awash in more consumer and market information than you ever imagined.

Benefits of Big Data for Marketers
Big Data can help you address a wide variety of marketing opportunities and challenges, facilitating information sharing to get tasks done more effectively and efficiently, and powering automated systems. Some other benefits of Big Data include:

Retain and upsell existing customers., the largest online retailer in the United States, uses its customer data to power its recommendation engine (the “you might be interested in this” offers that come up when you are on the site). Amazon at one time credited 35% of sales to this engine. Car manufacturers have been known to share website visitor behaviors, such as most-viewed car models and time spent on site with dealers after someone signs up for a test drive.
Identify new customers. Marketers have historically defined their audience and customer base in terms of demographics, such as age, marital status, and geography, based on the belief that people who share similar characteristics are likely to behave in the same way. Big Data empowers a move to one-to-one marketing based upon consumers’ actual behaviors and preferences. Beyond that, it can help you uncover new niche audiences by grouping people based on behaviors and interests rather than traditional demographics. ZestFinance, the credit scoring company, uses Big Data to weigh factors that typical credit score models miss. The company offers its analytics services to lenders so they can “better assess the credit risk of potential borrowers” and offer loans to people with low scores but a high likelihood of making their payments. The ZestFinance model prompted Quentin Hardy to post “Big Data for the Poor” on the NY Times blog site.
Reveal new marketing opportunities. You can also analyze Big Data to find patterns that yield new product segments or features and even new marketing opportunities. Online customer analysis by a leading appliance manufacturer revealed an opportunity to modify existing refrigerator product models to better serve customer needs.
Driving more profitable advertising. Audience data, coupled with recent advancements in real-time bidding, allows you to make more precise media decisions. With up-to-the-minute data about audiences’ behaviors and product interests, you can deliver highly relevant, one-to-one advertising messages. For example, if an airline knows that a customer has been researching a trip to Europe, it can deliver advertising creative that specifically references “Sale Fares to Europe.” This form of one-to-one advertising has been shown to dramatically improve campaign results. You can also put Big Data to work using predictive modeling to automatically optimize large-scale media buying activities.
Measure the impact of campaigns more accurately. Big Data can help you better understand how different media contribute to campaign performance, uncovering both opportunities and inefficiencies. By funneling data on impressions, clicks, conversions, social actions and more into attribution and media-mix models, you can learn how effectively each channel drives towards your desired objective, leading to better informed budget allocations. For example, an attribution analysis for a 360i client revealed $150MM of revenue that was directly driven by display media. Before the analysis, the brand had given display no credit for this revenue. The new understanding of display’s role in driving value prompted the advertiser to increase display-related revenues by 200%.

While this report focuses on how Big Data can be used to improve marketing, it is worth noting that big data has big implications for business beyond just the marketing department. A recent study found that data-driven decision making helps high-performing businesses achieve 5% greater productivity and 6% percent greater profits than their competitors.

Big Data Challenges
Capturing and analyzing Big Data isn’t easy. Marketers need to anticipate potential barriers, including:

Developing an infrastructure that supports Big Data. Because big data includes both enterprise data and, increasingly, data gathered outside the enterprise, you should understand at least at a high level what resources are needed — both technical and human — to collect, store, analyze and report on. That means lots of computing horsepower, disk space and technical know-how.
Finding ways to tie disparate data sources together. Big Data combines structured with unstructured data. Not only can it be difficult to figure out what it’s telling you, but you need to figure out how to tie it to an underlying data creator, such as an individual customer. For example, because tracking cookies are associated with a specific browser and device, you can’t easily tell if the person who spent two hours browsing your site yesterday on a PC is the one who used a mobile phone to find your store.
Aligning professionals from different parts of the enterprise. Analytics professionals speak an entirely different language than the business people they support. Companies need to find ways to help them communicate.
Upholding strict security and privacy protocols. It’s important to be careful with data, particularly personally identifiable information (PII). Law, regulations and standards vary widely by country. Becoming familiar with such mandates and guidelines is an important step in risk management.
Putting Big Data to Work
Regardless of your objective for using Big Data in marketing, consider these effective practices for approaching the process:

Appoint a chief data scientist or chief marketing scientist and develop a team. An important first step is to identify a central person in charge of data analysis— someone who has a broad understanding of what information is available to solve business problems within your organization. This person doesn’t necessarily need to sit in the C-Suite — it could be a head of insights or analytics — but he or she does need to have the authority to identify what data matters and determine the most effective way to analyze and measure it. While this leader must be technically savvy, he/she will not necessarily be the person who actually builds out your data infrastructure.
Educate your organization on the importance and availability of data. Another important role of the Chief Data Scientist is to educate the rest of the organization, especially the marketing department, on what data is available. The goal is to get people to think about how to use Big Data. Stimulating conversations that begin with “If I knew” is a particularly effective approach to discover opportunities in big data. For example:

If I knew what posts in social media were stimulating the most interest, how could I use them to improve my advertising creative?

If I knew which sports or special interests my customers were involved in, which events could I align with them?

If I knew when my customers were likely to walk past my door, what could I use to entice them to stop in?
Although Big Data is often tapped into initially to solve a specific business problem, simply knowing what data is available can help people think creatively about how to use it, yielding surprising and innovative results. With creative minds thinking about the data, your organization can go beyond just solving problems to uncovering new opportunities.
Identify tools to make Big Data more actionable. The right technologies are essential to get from messy unstructured data to business insight. A business intelligence (BI) solution is one of these — a very flexible reporting tool that enables marketers to slice and dice data with remarkable flexibility. Predictive analytics tools help you find the proverbial needle in the Big Data haystack, identifying traits and behaviors of individuals and groups and informing your marketing decisions. Technologies that automate marketing actions can produce effective results very quickly. For example, 360i built a technology that automates the results of predictive bidding for paid search — it can change bids and develop ad creative automatically across millions of keywords, something that would take a human being countless hours. Finally, data visualization translates complex data into visuals that can help marketers and their analytics kin overcome communication barriers and work more quickly towards those “aha” moments.
Take small steps for early, easy wins. Review the results of the “If I knew” exercise and match it to the data you have available or could easily acquire. Does anything pop? Also, review your organizational metrics to determine if they can be augmented by new data not currently in the mix. Are your Facebook fans part of your customer satisfaction index? Finally, put data from different sources side by side on a timeline. Do metrics seem to move together, and does positive sentiment in tweets relate to sales over time? These types of opportunities can become the low-hanging fruit that proves the concept and produces results that build your confidence in the use of Big Data, and reinforces your organization’s confidence in you.

Big Data discovery requires an open mind and an approach based on experimentation. Once you start seeing relationships in your data, you can start building the predictive models and marketing strategies that can help you capitalize on those relationships.

Big data is not an end in itself. Used wisely, it allows you to create a highly informed and effective marketing organization that can outflank your competitors and make you more relevant to your customers. Marketing data will continue to expand exponentially as more media is measured — via impressions, clicks, visits and social actions — and smart marketers must start to think about how they will leverage this data now.

In 2000, Billy Beane was the first general manager in major league baseball to use advanced analytics to fully exploit the data available to him. By 2003, a number of other teams had caught on. In 2004, the Boston Red Sox broke the curse of the Bambino and won the World Series for the first time in 86 years. They will publically acknowledge that they did so by hiring Billy Beane’s second in command — after offering Billy himself the largest general manager salary in baseball history — and implementing a rigorous advanced analytics program that capitalized on a richer set of data than anyone had used before. Today’s marketers need to emulate the Boston Red Sox. The alternative is to emulate the teams that lost to them in 2004.

Written by Kevin Geraghty, Vice President, Analytics & Reporting at 360i