1 1 Without Getting Buried by the Wave DataXu
2 2 Contents Without Getting Buried by the Wave Who s Afraid of Big Bad Data? Big Data? What I Really Need to Do is Understand My Customer! Big Data ROI in Action What to Look for in Your Big Data Partner What It Takes to Make Big Data Deliver You Can Harness the Big Data Revolution
3 Without Getting Buried by the Wave 3 If two companies use data with the same effectiveness but one can handle 15% of available data and one is stuck at 5%, who do you think will win? i This quote from Forrester s Brian Hopkins is a great wake up call. Big Data is big competitive advantage. New research from MIT Sloan School of Management Professor Erik Brynjolfsson puts a finer point on it. The research shows that companies which adopt data-driven decision making have productivity levels 5-6% higher. ii What is your ZMOT your zero moment of truth when your enterprise realized that Big Data and customer intelligence were more than mere buzz words? If it hasn t occurred yet, the time is now. Your customers are generating more digital data than ever. And the concomitant rise in digital marketing management platforms capable of harnessing this data means that your brand is ready for the revolution. In the next few pages, we ll gain a better understanding of the challenges of converting Big Data into business value. We ll explore the ways that Big Data delivers opportunities for value creation. We ll look at four real-world successes in harnessing Big Data. And we ll identify the core capabilities that any solution needs to offer in order to be able to convert Big Data into actionable intelligence. Who s Afraid of Big Bad Data? In 1,700 hours of interviews throughout 2011, CMOs from across the globe identified Big Data as the single biggest game changer they faced. iii IBM, the conductor of the survey, also asked which game changer CMOs felt least prepared to handle. Not surprisingly, the answer was Big Data. The volume, velocity, variety and variability of data are so great that marketers are in danger of being overwhelmed. At precisely the moment when real customer understanding seems so within reach, the avalanche of data makes it seem as far away as ever. In fact, Forrester estimates that firms effectively utilize less than 5% of available data. iv And it s getting worse. According to Gartner, enterprise data is expected to grow by 650% in the next five years v, much of that from the stunning increase in digital footprints created by consumers. Big Data? What I Really Need to Do is Understand My Customer! Enterprise data is expected to grow by 650% in the next five years. -Gartner What is Big Data anyway? EMC/IDC defines it this way: Big data is not a thing but instead a dynamic/activity that crosses many borders (and) Big data technologies describe a new generation of technologies and architectures, designed to economically extract value from very large volumes of a wide variety of data, by enabling high-velocity capture, discovery, and/or analysis. vi What makes this definition illuminating for marketers is that it describes Big Data in terms of the value it is able to generate. We would take it one step farther. If Big Data is the union of data and analysis, then customer intelligence is Big Data converted into action.
4 4 Big Data ROI in Action Enterprises that successfully convert Big Data into customer intelligence create value in multiple ways, according to a McKinsey Global Institute research paper vii. Here are four case studies that map to the ROI drivers they identify. In each case, marketers used DataXu s DX3 digital marketing management platform to make Big Data deliver on its promise of customer intelligence. Experimentation to discover needs, expose variability, and improve performance A top 3 computer printer manufacturer The printer company was able to shift mid-campaign to a more effective optimization strategy as a result of a happy accident. As a research component to a larger click optimization campaign, the platform ran a brand survey against different segments. The brand lift data was so compelling that the client shifted strategy and budget toward a brand campaign. Takeaway: Without the opportunity to experiment, marketers can continue down less than optimal paths, discovering months later that resources could have been better used. Segmenting populations to customize actions Properly analyzed, data allows marketers to identify increasingly granular customer segments. The era of 1:1 marketing is really upon us. A national retailer DataXu measured visitors to the home appliance pages of a major national retailer. After just a few days, the platform was able to provide new, dramatic insights that reshaped the retailer s marketing plan. For example, the retailer received a list of 3,000 third party data segments that indexed highly against the pixeled visitors. Organized in related groups, these microsegments offered the retailer a detailed roadmap for a campaign of precision audience buys with customized creative messages for each one. Takeaway: Without large volumes of tightly integrated 1st and 3rd party data, brands can t segment to uncover the kinds of correlations necessary to engage in customized marketing. Replacing/supporting human decision making with automated algorithms Algorithmic decisioning is not only necessary in the era of Big Data, it s better. And it frees staff to spend time on strategy, analytics and creative. Follow the money If anyone needs additional evidence that the Big Data revolution has arrived, look no further than the announcement from venture capital firm Accel Partners that it was launching Big Data Fund, a new $100 million initiative focused on identifying innovative entrepreneurs seeking to build category-defining companies at every layer of the Big Data stack. ix
5 5 A national insurer One of the United States top five national insurance companies used DataXu s cross-channel buying abilities to drive social engagement with the brand. On the open Web, DX3 s Active Analytics system identified several unlikely audiences that would be receptive to the insurer s message. As the Facebook campaign progressed, the system automatically shifted budget so that when the campaign concluded, three of the four newly targeted segments had outperformed the baseline. Takeaway: People and the data they generate don t respect artificial constructs like display vs. social. In fact, there are so much data moving so quickly that marketers must rely on automated systems to Replacing/supporting human decision making with automated algorithms Algorithmic decisioning is not only necessary in the era of Big Data, it s better. And it frees staff to spend time on strategy, analytics and creative. Scripps Networks Scripps, owner of networks like HGTV, DIY and the Food Network, is a leader in the home, food, travel and entertainment media categories. As both a publisher and advertiser, Scripps uses DX3 to solve a unique set of opportunities and challenges. When the platform started accumulating surprising data from its Paula Deen home cooking show, the Interactive team was able to expand its promotion to these new, receptive viewers, and packaged this insight to seek out new advertisers for its inventory. Takeaway: Without the unified intelligence made possible by a single-vendor integrated platform, marketers sacrifice the chance to discover non-obvious insights and incremental demand. What to Look for in Your Big Data Partner Before you write the RFP, before you start comparing feeds and speeds, step back and consider the qualities you want in a partner who will be so intimately tied to your business. That s what a Big Data vendor is a partner who is incredibly close to your customers. Transparency Simplicity Neutrality
6 6 Transparency Simplicity Neutrality There are two ways vendors can make money in marketing technology. They can exploit the appalling complexity of the landscape to hide the true cost of working with them. Or they can provide an honest accounting of costs and performance. Don t just ask the hard questions. Make your potential partner prove their answers. Digital marketing is complicated but advances in decision science have eroded the need for complexity. Find a partner with the breadth of capabilities that can reduce vendor proliferation. Your CFO will appreciate the lower total cost of ownership, and your team will be thrilled at how much wasted time and frustration you ve eliminated from their lives. Vendors come with a lot of agendas. Does the platform have to serve both publisher and brand interests at the same time? Does the vendor own the data it s trying to sell you? Pick a vendor the way you would pick a referee. Demand one whose goals are 100% aligned with yours. What It Takes to Make Big Data Deliver What does the ideal marketing platform have under the hood? Based on DataXu s experience, we believe it needs the following core capabilities in order to harness Big Data and deliver profitable customers and customer intelligence. Audience Management: Find profitable customers Effective platforms are those that can analyze large datasets from disparate sources and synthesize them into actionable segments. Marketers are ingesting data from their websites, ecommerce platforms, offline sales, third party providers and more. Without the ability to tap Data Management Platform (DMP) functionality, marketers will leave a lot of value on the Big Data table. Inventory Management: Decide where to communicate with them Marketers will never be able to become truly efficient if they have to rely on different vendors to find their customers on each of the screens they use during the course of a day. Look for a platform that can touch your customers across as many digital channels as possible. It will reduce the total cost to your marketing organization. Paging Dr. Hadoop If you ve spent any time looking at Big Data, you ve stumbled on the word Hadoop. Hadoop is an open source project that was initially developed to solve the data problems associated with indexing the Web for search engines. Hadoop can be applied to any problem involving large unstructured datasets like digital marketing. But enterprise IT analyst Dan Woods argues that CXOs should forget about the word for now: the question is not Should we use Hadoop? but What use should we make of data, big and small, and how am I going to be the leader in using data to help my business? x
7 Campaign Management: Execute that communication CMOs should look for platforms that are as good at discovering latent or incremental demand as they are at communicating with prospects predisposed to your brand. These Demand Side Platform (DSP) features should be easy to master, so that the CMO has the option of going self-serve, of taking the platform in-house. And the platform provider should have a service team that is responsive to their needs. Reporting and Insights: Understand the success of that communication Spitting out spreadsheets that must then be reconciled by in-house teams just won t scale in the age of Big Data. Expect a platform with reporting that is easy to understand and easy to work with, but also one with the sophistication to uncover insights that truly surprise you. Automation: Implement insights at the impression level Winning CMOs master Big Data faster than their competitors and for that, you need more than the passive reporting you get from most vendors. You need a platform that is able to use Big Data-driven customer intelligence while campaigns are in market, not after the fact. That means you need to demand a platform that has proven itself capable of making decisions and learning in real-time. We call that ability Active Analytics and it does not come easy. Be skeptical of vendors that claim it. Demand-Side Platforms For the uninitiated, a Demand-Side Platform or DSP is a technology that bids for digital media, impression by impression, in real time. This method of media buying is a replacement for the ad network arbitrage model (buy lower from publishers, sell higher to advertisers). First-generation platforms used rulesbased bidding. Newer platforms employ automated, algorithmic approaches that adjust bids in real time based on the performance of prior impressions. DSPs typically buy on ad exchanges which offer publisher inventory up for real-time bidding. 7 You Can Harness the Big Data Revolution The Big Data opportunity is large, but marketers should not underestimate what it takes to exploit it. In fact, McKinsey s research describes four levels of maturity in the enterprise s ability to profit from Big Data. 1. Digitizing and structuring the data 2. Making the data available 3. Applying basic analytics 4. Applying advanced analytics McKinsey elaborates on the fourth level automated algorithms and realtime data analysis that often can create radical new business insight and models...allow new levels of experimentation to develop optimal approaches to targeting customers. viii Most enterprises look at step four and are daunted: I ve spent millions on data collection and I have a huge team picking at it. But I can t get to the next level! With the right partner, the Big Data deluge can be channeled and harnessed. That partner can not only help you gain valuable insight into your best customers, it can help you find more of them, even as it reduces the total cost of ownership of your marketing programs. The right partner can Data Management DMPs are aggregators of data used by marketers to find customers and improve targeting. DMPs typically offer cookie pools organized into demographic, behavioral and psychographic segments. With the advent of Big Data and the ability to ingest 1st-party information including customer relationship management (CRM) data, these platforms have become increasingly important tools to help marketers organize customers and prospects.
8 8 Sources i shape_your_markets_next_big_winners ii iii iv v vi vii https://www.mckinseyquarterly.com/are_you_ready_for_the_era_of_big_data_2864 viii https://www.mckinseyquarterly.com/are_you_ready_for_the_era_of_big_data_2864 ix x your-ceo/2/