Tips & Tricks
Your SlideShare is downloading.
From information to insights: understanding Big Data online
Like this document? Why not share!
How to Assess Big Data Needs for Yo...
by Dell Enterprise
Unlocking big data
by Orchestrate Techn...
Email sent successfully!
Show related SlideShares at end
From information to insights: understanding Big Data online
The Marketing Distillery
Jul 03, 2013
Comment goes here.
12 hours ago
Are you sure you want to
Your message goes here
Be the first to comment
Be the first to like this
Number of Embeds
Flagged as inappropriate
Flag as inappropriate
No notes for slide
Transcript of "From information to insights: understanding Big Data online"
1. From Information to Insights: Understanding Big Data Online runner female likes to read uses Twitter uses Google+ Likes your brand tm
Copyright © 2013 Janrain, Inc. All rights reserved. Understanding Big Data Online | www.janrain.com | 888.563.3082 PG2 Table of Contents 3 Introduction 4 Evaluating the Online Consumer Data Landscape Social Network Stream Data . . . . . . . . . . . . . . . . . .4 Data Collected from Registration + Social Proﬁle Data . . . .7 Transaction Data . . . . . . . . . . . . . . . . . . . . . . . .8 Clickstream and Third-Party Data . . . . . . . . . . . . . . .9 10 How to Collect, Store, and Manage Online User Data Data Collected from Registration + Social Proﬁle Data . . . 12 Social Network Stream Data . . . . . . . . . . . . . . . . . 12 14 Best Practices to Take Advantage of Online Big Data Email Segmentation. . . . . . . . . . . . . . . . . . . . . . 14 Product Recommendations & Content Personalization . . . 15 Ad Targeting. . . . . . . . . . . . . . . . . . . . . . . . . . 18 20 Conclusion
Copyright © 2013 Janrain, Inc. All rights reserved. Understanding Big Data Online | www.janrain.com | 888.563.3082 PG3 <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Never before have digital marketers been blessed with such immense quantities of online customer data at our disposal, yet been so incapacitated by how to use it. From personal information collected at registration to transaction history, clickstream data, social network streams and third-party data, marketers are sitting on a veritable gold mine of customer intelligence – one that remains largely untapped. This massive volume of consumer information has led to the coining and popularization of the term “big data”. But the advent of big data has yet to be followed by established practices for managing and using such data. In short, strategy and execution have not accelerated at the pace of the technology, which has led to more questions than answers. How can you extract value from the ever-increasing volume of customer-related information? How do you transform information into insight? These are the questions keeping marketers up at night. And they are hindering our ability to truly connect with customers and justify return on marketing investment. This paper will explore the common sources of online data and provide practical advice for accelerating the collection, storage and most importantly, use of this data to drive business objectives. Introduction
Copyright © 2013 Janrain, Inc. All rights reserved. Understanding Big Data Online | www.janrain.com | 888.563.3082 PG4 <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm The Consumer Data Landscape Evaluating the Online Consumer Data Landscape The biggest misnomer about big data is that it only applies to large data sets. But as Gartner pointed out way back in a 2001 report, massive volume is only one deﬁning element of big data. Variety and velocity also play signiﬁcant roles. Variety refers to the diversity of data formats that exist across multiple sources, which traditional database solutions often are poorly equipped to handle. These data types may include combinations of text strings, images, web logs, documents, numeric data, plurals, unstructured blobs, Boolean data and many others. Marketers and IT wrestle with inconsistent and unstandardized data on a constant basis. Without the tools to make sense of the data, they are ﬁghting a losing battle in their attempts to understand and leverage this data to achieve corporate objectives. Velocity refers to the rate of change for customer data and how quickly it must be utilized in order to be valuable. Clickstream data is a perfect example. The pace at which technology systems are accumulating information on consumer online browsing habits is staggering, yet most marketers lack the infrastructure to capture and utilize this information. Without near real-time utilization of clickstream data, decay rates set in and targeting efforts become an exercise in futility. Is it really worth it to target consumers based on behaviors logged weeks or months ago? Much like attempting to collect rainwater with a colander, traditional database solutions simply are not built to handle frequently changing data sets. Have digital marketers ever been faced with such a stark dichotomy? On one hand, there’s a tremendous opportunity to leverage the volume, variety and velocity of data out there to connect with consumers. But on the other hand, immense confusion exists about what big data is and how to utilize it. To shed some light, we have prepared a primer on a few common online data sources and how they are evolving. Social Network Stream Data There is a tremendous amount of user- generated content available on social networks. In fact, people shared more than 8 billion pieces of content on Facebook each day in July 2012, up from 4 billion per day the previous year. This information can deliver near real-time consumer insights if marketers can ﬁnd a way to make it useful. Purchase intent, brand advocacy and customer service issues can all be gleaned by mining
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG5Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm social network data streams. But doing so isn’t easy – these streams contain massive quantities of unstructured data that lack a pre-deﬁned or consistent model, and unearthing insights from them is a formidable and costly task. Accurate interpretation of sentiment is one challenge marketers face when using technology algorithms to analyze social network data streams. Consider parallel tweets about the iPhone 4S from two different people: Both tweets utilize the same matching keywords - the object of both is the iPhone 4S, and the adjective used to describe the phone in both is “sick”. Most social media monitoring technologies would have trouble deciphering the semantic nuance of these tweets. But any human with a cursory understanding of the English language and Generation-Y vernacular would have few problems interpreting their very different meanings. One of these Twitter users should be identiﬁed as a brand advocate, while the other should be characterized as a candidate for customer service. Consider a second challenge faced by marketers when seeking to mine the social stream. This tweet clearly indicates purchase intent on the part of the Twitter user. But how many data mining services could easily parse the semantics of this tweet to determine intent to purchase a replacement phone? While mining the Twitter stream for marketing insights at scale has remained a tough nut to crack, Facebook, by contrast, has actually laid the necessary groundwork to make such analysis possible from its own data sets. There was a time when Facebook proﬁles were full of unstructured data ﬁelds. Facebook users could type in their favorite music, books or movies into a text area ﬁeld, with each distinct interest separated by a comma. The loose validation rules and free form text ﬁelds allowed for typos (“Sienfeld” instead of “Seinfeld”) as well as vague, decoupled data ﬁelds (such as a declared interest in “making cool things”). The Consumer Data Landscape
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG6Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Then, in April 2010, Facebook launched an update to proﬁles. The impact of this update slipped under the radar, but it drastically enhanced the ability for Facebook as well as savvy marketers to extract value from the data within a proﬁle. Facebook essentially indexed all hobbies, music artists, books, movies, television shows and other attributes, and created public community pages for each. Facebook then mapped a user’s declared interests and other attributes to these pages, which enforced a more rigid structure for the data. The change helped standardize Facebook’s data set on users, and eliminated the possibility of typos, erratic free form entries, or unstructured proﬁle data. As evidence of its new data validation rules, if a Facebook user currently attempts to declare an interest that is not indexed as a public community page, an error message appears: The change was a colossal step forward toward improving the integrity and usability of social data contained with a Facebook proﬁle. Facebook’s commitment to transforming the contents of its users’ social proﬁles from unstructured data to consistent, structured data, in conjunction with the advent of tools that enable brands to collect this data from a Facebook proﬁle, has uncovered exciting new possibilities for marketers. More on this later. So, if 2010 was the year that Facebook tackled its unstructured data problem within its user proﬁles, 2011 was the year it sought to do the same for the Facebook news feed stream. Similar to Twitter, the deluge of posts, updates and shared content represents a gold mine for marketers who can ﬁnd a way to establish meaning from the clutter. Facebook’s launch of Open Graph is expected to be the ticket to enable such insight. The Consumer Data Landscape
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG7Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm The cornerstone of Facebook’s Open Graph is the ability to fuel contextual social sharing by enabling object-verb associations within a shared message. Rather than millions of Facebook users “liking” products and content, Open Graph goes a layer deeper by enabling websites to structure a more meaningful verb with its product or content portfolio. Instead of simply liking a product and sharing it back to their news feed, consumers can now specify that they want a product, or own a product. This context is invaluable for retailers seeking to enable effective real-time intent targeting and segmentation. Similarly, digital publishers can determine whether a reader has read an article versus commented on the article, helping them map levels of subscriber engagement to their content portfolio. Clearly, Open Graph gives users a more contextual sharing experience and brands a better opportunity to gain relevant exposure within the news feed. But aside from achieving those goals, what’s in it for Facebook? Think about all of the data Facebook is now collecting from users as they navigate across the web. Open Graph-enabled websites are passing large volumes of semi-structured data back to Facebook in real-time. In essence, Facebook has the ability to begin aggregating clickstream data from its users as they traverse the web – articles they read, products they want, recipes they cook, and so on. Over time, Open Graph could enhance Facebook’s ability to monetize its data by functioning as a plumbing system of real-time information from the rest of the web. For brands to tap into this information stream and reap the beneﬁts of deeper user intelligence, they will ﬁrst need to employ tools that will collect this semi- structured data and extract meaning from it. Data Collected from Registration + Social Proﬁle Data Deriving insights from ﬁrst-party data assets is the single most important factor driving increased investment in marketing data, according to a 2012 report jointly published by Winterberry Group and Interactive Advertising Bureau. And more than any other characteristics, marketers assess the utility of a data set based on its quality and accuracy, followed by its recency or “freshness”. Purchase Intent Customer Loyalty Target The Consumer Data Landscape
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG8Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm The Consumer Data Landscape Yet, despite these priorities, marketers are relying on a paradigm for acquiring ﬁrst-party data, known as traditional registration, which undermines the integrity as well as recency of user proﬁle data. A 2011 report commissioned by Janrain showed that 88% of consumers admitted to providing false information when registering at a website the traditional way. Compounding matters, for the 12% who actually provide accurate information at registration, such data exists in a time warp. Once a consumer has registered on your site, what is her incentive to keep her proﬁle up to date? If her email address changes, is she going to let you know? How about her location? Or relationship status? First-party data collected from users via traditional registration methods is littered with ﬂaws that inhibit its value for marketers. What’s the solution? Consumers already actively maintain identities on social networks and email providers such as Facebook, Twitter, Google and Yahoo!, and 77% would prefer to use one of those existing identities to register on sites. Social login lets people securely and easily sign- up on a site within just two clicks using an existing social identity. During this process, consumers can choose to share demographic and psychographic information from their social proﬁle with a brand site. Marketers can gain a more sophisticated understanding of their consumers by leveraging the proﬁle data that people already maintain on their social networks. Social proﬁle data includes not only basic demographics such as name, age, gender, geography and email address, but also deeper psychographic information such as interests, marital status, political views, hobbies and friends. And because the proﬁle information that users maintain on their social networks is transparent to friends, family and coworkers, it is more likely to be current and accurate than personal data that consumers may supply during a traditional registration process. “77% of users would prefer to use an existing identity to register on sites.
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG9Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm The Consumer Data Landscape Transaction Data Let’s be clear – transaction data, which describes change as a result of an event, certainly doesn’t scale at the level of clickstream data. The number of consumers who complete a purchase or add a subscription on a brand’s site will never compose more than a fraction of total visitors to that site. But transaction data’s value to marketers cannot be understated. Past purchase history often serves as a basis for merchandising, by enabling marketers to model purchases to complementary offerings within a product catalog. Additionally, transaction data enables marketers to identify current customers from suspects, and use such intelligence to create targeted segments. Clickstream and Third-Party Data Clickstream data is “big” due to its sheer volume and velocity. There’s a lot of stuff happening online, and it’s happening very fast. Technology companies have been analyzing web usage since the 1990’s, and the scale of data amassed has placed their services in high demand. The power brokers of the clickstream are data management platforms (DMPs) and data aggregation services. Put simply, these technologies use browser cookies to track what you do, where you do it and when you did it on the web. Thankfully, they don’t yet explain why we do it, lest we want computers deciphering the web’s obsession with cats! This data is mostly anonymous, meaning it has been stripped of personally identiﬁable information (PII). Data management platforms and personalization engines aggregate third- party clickstream data and develop clustered audience segments based on users who share similar characteristics – males in San Francisco, CA who like surﬁng, for example. Brands can use data management platforms to build audience segments that inform advertising campaigns. Once marketers develop an understanding of how to create the right audience segments based on business intelligence, there are ways to augment third-party data sets and make them more relevant. Using online matching, marketers can combine explicit ﬁrst-party data they collect on their own with third-party data to create better segments. As an example, by passing transaction data through to a DMP, a big box retailer can better understand the relationship between gadget enthusiasts and purchase behavior on its site.
Copyright © 2013 Janrain, Inc. All rights reserved. Understanding Big Data Online | www.janrain.com | 888.563.3082 PG10 <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Managing Online User Data How to Collect, Store, and Manage Online User Data Marketers have created fences that must be torn down. According to the Winterberry Group/IAB report, two of the most signiﬁcant challenges inhibiting marketers from investing in big data are: 1. An absence of tools to unify data silos. 2. The lack of a cohesive strategy to utilize the data. As described above, marketers are overwhelmed with data. It comes from different sources, it’s stored in different systems, and different tools are used to analyze it. In short, digital marketers and IT professionals lack “a single version of the truth”, which makes it hard to develop any version of a strategy. About the big data transformation, the study claims: “Most other marketers – saddled with legacy technology platforms, depleted of expertise by years of underinvestment and structured only to support “traditional” approaches to data usage – are ﬁnding they’re woefully unprepared for this transformation. ”
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG11Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm You see, even the best laid strategies and plans can be undone by walled gardens. Before marketers can start utilizing big data, they need the right infrastructure in place to fuse silos and build a uniﬁed view of a customer. Consider, for example, the average retail company within Internet Retailer’s Top 500 list. Most (72%, to be precise) use an in-house CRM system to store customer data – things like name, email address, geographic data, and username and password. But customer information also gets collected and stored in a dedicated eCommerce system when customers purchase products on the site. A slight majority (53%) also use in-house systems to manage these eCommerce transactions. Then there is the payment system, which must collect and store pieces of customer information as well, the email marketing solution to enable sending of offers to shoppers, the fulﬁllment system, which consumes a shipping address, the personalization engine to aggregate clickstream data, and the ratings & reviews solution, which often collects portions of a shopper proﬁle for community engagement. Now, imagine trying to build a 360-degree view of a customer by integrating data elements across each of these disparate technologies. Many vendor technologies don’t integrate with other systems, and those retailers or publishers using in-house systems are essentially resigned to lengthy and expensive custom development work to build integrations. This makes it exceedingly difficult to extract the insights and build segments for effective marketing campaigns. Marketers are focused on building deeper relationships with customers, but without an easy method of sharing data between systems, that vision becomes hazy. There has to be a better way. As cloud-based technologies and software-as- a-service offerings have matured, integrations across vendors are becoming more common. Whatever user management solution you employ as your database of record to acquire user data and understand your customers, make sure it provides seamless connectivity to share up-to-date proﬁle information between applications. Data-layer connectors enable marketers to easily map proﬁle data ﬁelds from one system to the other, and create real-time, batch or on-demand ﬂows of user data between platforms. Your email marketing system, for example, shouldn’t have a different idea of who a particular customer is than your CRM. Digital marketers should choose technology systems wisely on the basis of how well they play within a company’s existing technology stack. Let’s again dive into different categories of ﬁrst- party data to uncover methods for collection and storage. Managing Online User Data
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG12Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Data Collected from Registration + Social Proﬁle Data Earlier, we described how social login unlocks the door to obtaining more accurate information about your consumers and building richer proﬁles. Because the proﬁle data collected via social login includes demographics, rich psychographic information, interests and friends lists, social proﬁle data is more comprehensive than other ﬁrst-party data sources, and more reliable and accurate than the clickstream or third-party data. Once marketers have access to social proﬁle data, the next step is to store and utilize it. With some effort, legacy databases can be retroﬁtted through schema modiﬁcations to store social proﬁle data. Or, marketers can choose an off the shelf social proﬁle storage solution that automatically captures this data and pre-integrates with email, content personalization, eCommerce and other technology systems. If choosing an off the shelf solution, it’s critical to grasp how the technology handles schema for proﬁle data. When it comes to social proﬁle data, the ability to enforce structure and rules on such data, through schema, typically results in much better data quality and makes it faster and easier for marketers to build granular customer segments and feed user proﬁle data into different technology systems (email, personalization) for campaigns and programs. Social Network Stream Data We talked above about the wealth of real-time updates pouring into social feeds. In order to mine such data for semantics, you ﬁrst need to collect it. Sounds obvious, right? At ﬁrst glance, social listening technologies are a solid place to begin your great big social stream data excavation project. Most of these tools allow marketers to export data from social feeds into common formats such as .CSV. The exported ﬁles typically include Twitter handles or Facebook identiﬁers to help uniquely identify consumers. Another option is to gather such data either by using APIs offered by social networks, or social login vendors that manage these APIs. Facebook, for example, offers scopes and permissions that enable brands to access a Managing Online User Data
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG13Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm list of all posts within a user’s news feed when that consumer signs in to a brand’s site using a Facebook identity. Searches and queries can then be performed against that data to mine for keywords that would provide customer intelligence or indicate intent. Similarly, brands can access a list of a user’s mobile check-ins with permission in order to identify places that person has visited and mine for targeting opportunities. Now, determining how to store all of this data is far from a trivial task. The volume and velocity of data in social feeds makes it a poor ﬁt for traditional database software. Beyond social listening tools, explore modern proﬁle storage solutions that utilize a social data model. A social data model ensures that the architecture and schema of the database can handle ﬁelds of unstructured or semi- structured data while enabling ﬁltering and high-performing queries. Managing Online User Data
Copyright © 2013 Janrain, Inc. All rights reserved. Understanding Big Data Online | www.janrain.com | 888.563.3082 PG14 <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Best Practices to Utilize Data Best Practices to Take Advantage of Online Big Data As mentioned earlier, technology has outpaced strategy in the evolution of the digital era. We’ve discussed the composition of big data and how it has become more accessible and functional for marketers. Once the technology is in place to access and store such data, marketers need the skills to actually take advantage of it. Historically, we’ve turned to the clickstream and information contained within cookies as a go-to data source for targeting programs. While clickstream data can deliver an adequate understanding of consumer behavior, past behavior isn’t always indicative of future intent. Digital marketers need to evolve their targeting efforts from what people have done to how they think. Social proﬁle data holds the key to this insight. Here are several ways to employ big data insights into your digital marketing campaigns, while blending social proﬁle information with other sources of consumer intelligence. Email Segmentation Email marketing is low-hanging fruit for digital marketers seeking to apply big data insights to connect with their consumers. Compared with online advertising or content personalization, the skills and barriers to entry required to execute effective email campaigns are much lower than with other marketing programs. Unlike advertising and content personalization, whose complex algorithms are heavily reliant upon third-party data, successful email campaigns can be executed exclusively using data a brand owns – registration and transaction information. B2B marketers have been doing this for years. Using marketing automation solutions, B2B marketers create customer segments based on industry, job function, and stage in the buying cycle. Prospects within these target segments then receive a sequence of triggered emails that are designed to induce brand recall, purchase intent, or loyalty.
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG15Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Certainly, B2C marketers have little use for a potential buyer’s job role or industry. But using social login, marketers can instantly gain permission-based access to the information they care about - rich demographic data and interests straight from a consumer’s social network proﬁle. This data set not only includes a pre-veriﬁed email address, name, location and birth date, but also relationship status, political views, hobbies, favorite books, music, movies and television shows. The key is to create micro-segments of consumers who share similar demographic or psychographic characteristics and use them as the basis for targeting. When stored in a ﬂexible user management system and combined with transaction data from an eCommerce or subscription platform, marketers can send highly relevant offers to consumers via email. For example, let’s say that a registered user in your database has declared an interest in running. Perhaps some of your newest running apparel might be an appropriate recommendation for this consumer. But what if you could match that declared interest with her past purchase history, which indicated that she already purchased a tank top and running shorts from your company. Based on this insight, a more appropriate recommendation might be that popular pair of running shoes in your product catalog. Research indicates that targeted emails containing personalized content and offers enjoy a nearly 4X greater click-through rate than generic email offers. By selecting the right infrastructure to store and leverage social proﬁles and consumer data, and taking the time to build intelligent segments, marketers can dramatically improve email marketing ROI. Ensure that your email service provider integrates tightly with your user management database, and that seamless sharing and transfer of consumer data across both applications is easily achievable. Product Recommendations & Content Personalization Online retailers and digital publishers traditionally have relied on one method to inform product recommendations and content personalization – predictive modeling based on clickstream data or transaction data. Amazon. com pioneered the use of clickstream and transaction data for personalization years ago by recommending products based on your browsing behavior and purchase history. How does the technology actually work? Retailers and publishers generally outsource it to specialized personalization engines. These vendors aggregate data from multiple sources, including browsing behavior or purchase history, and utilize modeling techniques to group people into clusters based on those attributes and behavior patterns. These audience segments form the basis of behavioral targeting and personalization, but they are generated from archetypes or personas and not uniquely tied to real consumer identities. Best Practices to Utilize Data
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG16Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm From a privacy perspective, this is a good thing. Unless you’ve opted in to sharing personal information with a brand, do you really want that company or a data mining service tracking behaviors that are uniquely tied to your identity? The limitations with the most current methods of personalization is they often fall short at providing relevant and reliable recommendations for consumers, and the data sources are not always stable or actionable for sustained use in marketing programs. Personalization engines dynamically serve content as part of the presentation layer on a site, but brands often lack access to the underlying consumer data that informs a recommendation. And clickstream data is only as persistent as the tracking cookie placed in a user’s web browser upon her ﬁrst visit to a site. Once that cookie is removed, a brand relying on clickstream behavior for product or content recommendations must start over from scratch. So, the clickstream is not dependable due to its fragility within a browser cookie. Nor is transaction data universally reliable as a predictor of intent, due to gift and experiential purchase behavior. That power tool purchase I made as a gift for Father’s Day does not indicate that I want to be served recommendations for similar products – power tools are not aligned with my interests. Similarly, recommendations for travel guides to Costa Rica long after my trip ended are not likely to induce future purchase intent. Here’s the other challenge with transaction data for retailers – purchase conversion rates for ﬁrst-time eCommerce site visitors hover at about 3-4%, meaning that most of your visitors have never purchased a product from your site. Without a historical record of past purchases to inform recommendations, retailers lack data scale and are left in the dark. Best Practices to Utilize Data Personalization
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG17Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Best Practices to Utilize Data If clickstream data is unreliable and transaction data isn’t scalable across your audience or the best predictor of future purchases or content interests, then what is? A few retailers and publishers ahead of the curve in this space are hedging their bets on social data, speciﬁcally leveraging a consumer’s interests, real-time sentiments and social graph to inform recommendations. As mentioned above, audience segments used in personalization algorithms are generated from archetypes and not uniquely tied to real consumer identities. But when a consumer explicitly shares her social proﬁle data with you, the paradigm changes. Now you actually know the potential customer you’re targeting. Imagine if REI knew that a shopper was really interested in bicycling and snow skiing from the moment she connected a social identity on its site, and could tailor product recommendations to prevent her from needing to wade through troves of camping or climbing gear. Social proﬁle data makes this possible by providing brands with opt-in access to a consumer’s interest graph – information such as likes, hobbies and real-time status updates, all of which can be appended to existing data structures to augment personalization efforts. But social proﬁle data isn’t limited to interests. Retailers such as Sears, Levi’s and Etsy are starting to leverage a consumer’s friends to create social shopping experiences. Sears.com lets customers share their list of friends from Facebook with the site, and then recommends gift ideas based on the birthdays and interests of the social graph: For digital publishers, the same principles of personalization can and should be applied. Consider the “product” in this scenario to be your site content. Serving up articles to readers based on demographics and interests they have declared in their social proﬁle, or based on their friends’ activities, is a great way to increase subscriptions, time on site, commenting activity, ad impressions and CPM yield. Gift Suggestions
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG18Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Ad Targeting Complicated by countless data sources, intricate delivery algorithms and privacy concerns, success in online advertising doesn’t come easy. Digital marketers lean heavily on advertising networks and data brokers to provide the data and technology required for audience segmentation. Such information is collected almost exclusively via browser cookies. In May of 2012, Microsoft struck a blow to publishers and online advertisers by announcing its decision to turn on the “do not track” speciﬁcation by default within Internet Explorer 10. The do not track ﬂag enables consumers to opt out of behavioral targeting. Do not track is predicated on voluntary compliance from ad networks and publishers, and at this point, while advertising groups object to Microsoft’s decision, they are offering to honor the standard. With Internet Explorer enjoying 32% share of global web browser usage, Microsoft’s decision is likely to reverberate throughout the industry. Despite a historical reliance on tracking cookies because of their ability to reach scale, advertisers and publishers may soon need to resort to other data sources in order to achieve ROI from personalized ad units. First-party data collected from registration could suddenly become inﬁnitely more valuable for advertisers. Because not all visitors to a site will choose to self-identify to a site via registration, ﬁrst- party data collected at account creation or via social login will never reach the scale of third- party data assembled via cookies. But there is an opportunity to apply this data within ad campaigns to augment consumer intelligence, improve click-through rates, lower cost of acquisition, and generate more incremental revenue per impression. The key is to understand data governance and deploy a user management system that is capable of integrating with data management platforms. As the Winterberry Group and IAB study observes: “Many are coming to see marketing data governance – deﬁning the ‘rules of the road’ for assigning distinct data sources to different promotional tasks – as equally important. ” Why is this important? Facebook and LinkedIn, for example, maintain strict guidelines that restrict how social proﬁle data can be used in the context of advertising. As marketers collect social proﬁle data and employ progressive proﬁling tactics to build richer proﬁles on their consumers by gradually requesting more information over time, it is critical that their database helps them understand the source of a data set and how such data can be used for ad targeting. Best Practices to Utilize Data
Understanding Big Data Online | www.janrain.com | 888.563.3082 PG19Copyright © 2013 Janrain, Inc. All rights reserved. <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm When handling social proﬁle data from these networks, a database solution should assist data governance by helping marketers and IT strip out personally identiﬁable information from a data sample where necessary before transferring it to a data management platform. For example, for the social networks with stricter terms of service about the use of social data for ad targeting, a comprehensive user management solution can automatically categorize stored user data based on its source, and prevent its circulation to data management or targeting systems. For all other registration and proﬁle data stored within these systems, a common strategy is to create lookalike audience segments by modeling explicit ﬁrst-party information against other aggregated data. Other networks such as Twitter are less restrictive about the use of interests from a consumer’s proﬁle for ad targeting. This is great news for marketers seeking to build interest-based audience segments without the use of clickstream data. When social login is offered, marketers gain permission-based access to a consumer’s Twitter data. Querying that consumer’s Twitter bio and tweet stream for keywords (such as politics, running or other hobbies) delivers granular, interest- based audience segments without the use of clickstream data. In short, Do Not Track doesn’t need to signal the end of psychographic targeting. While the future of online advertising (when using more reliable permission-based proﬁle data) may result in lower volumes of targetable consumers, cost of acquisition per user and return on investment ultimately will improve. Best Practices to Utilize Data Romantic Honeymoon Getaways Male + Female Engaged Fargo, ND Proﬁle Attributes Targeted Ad
Copyright © 2013 Janrain, Inc. All rights reserved. Understanding Big Data Online | www.janrain.com | 888.563.3082 PG20 <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm Conclusion In the vision of the future web outlined by Tim Berners-Lee, technology will evolve to understand the meaning, or semantics, of information on the World Wide Web. When it comes to collecting insight from the torrent of online consumer data, this hasn’t happened quite yet, partly because computers are not smart enough but also because we’re constrained by data silos and lack a cohesive strategy. But with time, the data silos will disappear, and advanced tools to mine social network streams and utilize social proﬁle data will become a necessity for brands. As they do, marketers will gain the technology interoperability and insight needed to truly harness the power of big data. Will you be ready?
Copyright © 2013 Janrain, Inc. All rights reserved. Understanding Big Data Online | www.janrain.com | 888.563.3082 PG21 <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> tm About Janrain The Janrain User Management Platform (JUMP) helps organizations succeed on the social web by providing leading technology to leverage the popularity of social networks and identities for user acquisition, engagement, and enhanced customer intelligence. Our solutions, including social login, social sharing, social proﬁle data collection and storage, access to the social graph, and digital strategy services, improve the effectiveness of online marketing initiatives for leading brands like Fox, Universal Music Group, Whole Foods, MTV, Purina, Avis and Dr Pepper. Founded in 2005, Janrain is based in Portland, Oregon. For more information, please call 1-888- 563-3082 or visit www.janrain.com and follow @janrain.