Mfarhanonline Social Media News: Ray Velez is the CTO at Razorfish , and the editor of new technology report, Razorfish 5 : Five Technologies That Will Transform Your Business. In this era of constant connectivity and tracking , we are creating huge amounts of data every second. Just like consumers, marketers are facing information overload. The big question is: How can we use this data to produce better, more relevant customer experiences? The data that consumers are creating is allowing marketers to listen and respond to their customers in ways never before possible. This data set resides in huge clusters of information that paint a picture about what is relevant to users. Data coming from media and advertising, including ad servers like Doubleclick and Microsoft Atlas, will tell us what creative and messages users have been exposed to, how many times, where have they have seen it and if they have clicked to find out more. Site analytics data, coming from web analytics tools such as Google, Omniture and Webtrends, tell us everything from how long people spent viewing content (in more detail than number of pageviews), how often they visit the content, and even what they are searching for and not finding. Finally, social data from Facebook and Twitter can tell us who is in a consumer's social graph and what brands and content they and their friends like. However, putting all of this data into a usable and effective format for brands takes some time. Fortunately, there is an industry-proven process that allows us do just that. Organizing Big Data Marketers can break down and manage this information: Distinguish the unique identifier across all the data sources. Connect ad cookie data with web analytics cookie data to build the profile of each unique identifier. Connect that profile with data already logged in from other sources, including profiles with Faceb! ook Twit ter IDs. Continue to build on this basic profile, adding new data from sources like Foursquare as they become available. After completing this process, we have access to trillions of rows and petabytes of data that traditional tools can't interpret. So what do we do now? Fortunately, we have new amazing technologies and tools at our disposal, thanks to some of the major technology companies. A few years back, Google shared its design for dealing with huge volumes of data in a paper called " Big Table ." Yahoo took this paper and wrote an open source implementation of it called Hadoop . Amazon Web Services customized Hadoop and called it Elastic MapReduce . As a result, any marketer can rent the same technologies that these big companies use on an as-needed basis to churn through huge amounts of data without a huge investment. This has dramatically increased a company's ability to personalize their experiences. While Hadoop and Elastic MapReduce have set the standard on how to scale and aggregate large data, there are still a couple of steps required to actually make sense of the data. At the highest level there are two approaches. One is to take the data and put it into a traditional data warehouse, which enables us to slice and dice data across facts (i.e., brand engagement) and dimensions (i.e., time). While this work is great and has been around for a while, it is often rigid and difficult to perform new types of analysis. Fortunately, there are new and exciting technologies enabling us to do warehouse-style analysis on raw data without the time and infrastructure the old model required. Technologies like Netezza and Aster Data enable working directly with the raw de-normalized Elastic MapReduce data without having to predefine and build traditional data warehousing cubes. When we applied this approach to one of our retail client's data, we got amazing results. We created about 36 personalization segments, and these segments were able to drive close to 2 million personalized messages. We wou! ld impor t about .5 trillion rows of data into a 100 machine cluster on Amazon Web Services to personalize our experiences. This all happened in about eight hours and culminated in about a 500% return on advertising spend. If we had used traditional technologies and methods, this would likely have been cost prohibitive, not to mention it would have taken weeks or months just to get all the servers up and ready. But since the cloud uses vast resources much more efficiently, we were able to only use what we needed. Making Big Data Work for You Working with these vast new troves of information and translating them to actionable profiles requires a data scientist — someone who is comfortable with the data constructs like cubes, pivot tables, etc., and can map them to the profiles or personas that are being defined by the marketing team. It’s not only about mapping the data to the marketing personas but helping to define those personas. So, how do marketers use this data and persona information to develop better web design experiences? A single view of any page on your brand's website assumes every visitor looks at your message or transaction the same way. Marketers must start thinking of homepages and websites as something that should be tailored for any number of profiles or segments. Lastly, your company's web development team needs to think about how to connect external data sources with the ability to personalize and tailor your sites. Should different segments have different search relevancy or be served different navigation? Leveraging tools such as FAST , Endeca , Google and others enable the creation of different search collections. Commerce engines such as ATG or Demandware or personalization tools from content management system products like Adobe, Tridion , or Autonomy can also be used to create more personalized experiences. In conclusion, there is a huge opportunity to listen to your customers and create relevant experiences on a scale never before possible. The data is there, and with ne! w thinki ng and cloud technologies, we can use it to drive richer experiences and better return on investment. Interested in more Business resources? Check out Mfarhanonline Explore , a new way to discover information on your favorite Mfarhanonline topics. Image courtesy of iStockphoto , Yakobchuk More About: business , data , MARKETING , tech For more Business & Marketing coverage: Follow Mfarhanonline Business & Marketing on Twitter Become a Fan on Facebook Subscribe to the Business & Marketing channel Download our free apps for Android , Mac , iPhone and iPad Social Media reviews series maintain by Mayya
0 comments :
Post a Comment