Unless you have been living under a rock the last few years, you probably came across the term big data at some point. Maybe it was around the water cooler with colleagues. Or, maybe (more than likely) you saw references to big data at an industry conference, tradeshow, or event. But, despite all the buzz on big data, most marketers still have more questions than answers. Well, we did to.
Here’s what you probably noticed:
- Definitions on big data are pretty nebulous
- Most discussions of big data are more esoteric than real
- References to big data include hardware, software, processes, and strategy
- As a marketer, it’s unclear if you should give a darn about big data
Last month Gleanster published a complimentary Deep Dive report on big data called “A Crash Course on Big Data . . . for Marketers“. This report provides a comprehensive overview of the various caveats and elements that shape and define what big data means and how it’s relevant to marketers.
So, if you were looking for a high level “give me the facts jack” marketing perspective on big data, here it is:
_________________________________________________
10 Things Every Marketer Should Know About Big Data
The Definition of Big Data
One widely accepted definition refers to big data as “data that is so large and complex it has outgrown the ability to aggregate and process it from a single database or repository“. Think about accessing all the different customer data in your organization across every marketing interaction – now layer in finance, operations, sales, and customer service data. That’s big data and it’s far too much data for one database. In theory, the ability to analyze massive volumes of data should provide additional analysis for data driven decisions that had previously been impossible to extrapolate.
Why is big data relevant to marketing?
It’s your fault. Marketers are largely in the driver’s seat when it comes to adopting new technologies that generate large volumes of data. At the same time, marketers stand to gain the most from having access to actionable data that can ultimately impact top line growth. When you think about the massive volume of data your organization collects on a daily basis, and how much of that data goes largely untapped, it’s not hard to understand why the term “big data” became popular. In the context of marketing, every organization is sitting on a massive amount of customer data. From web analytics data, to customer purchase data, to social media interactions, to customer segments, behavioral attributes and demographics – and all of that grows exponentially over time with repeat purchases from loyal customers, new products, etc.
Is big data hardware, software, or a concept?
Big Data is all of the above; software, hardware, strategy, services, etc. Big data defines a holistic approach for making large volumes of data accessible for analysis. From a hardware perspective, there are new hardware solutions (like Oracle Exadata and Violin-Memory) designed to process large volumes of information and/or store it in-memory for rapid access (it turns out there are limitations to storing information on spinning disks (hard-drives). By moving the data into large in-memory storage (think USB flash memory in the form of a refrigerator) data can be processed much faster and in real-time. Here’s what matters for marketers. Big data solutions are all about making massive volumes of data available for analysis by business users. In a big data solution information is real-time, the interface is intuitive, and there are new forms of data visualization and analysis. It’s a powerful combination of hardware, software, and of course IT strategy that makes large volumes of customer data accessible for analysis. That means, marketers won’t have to go to IT and submit a ticket to process a query from the database (which could take days). Big data circumvents that entire process and puts analysis in the hands of marketers.
Exactly how big is big data?
Put differently “How much wood would a woodchuck chuck if a woodchuck could chuck wood.” Big data is big, but there are limits to how much data exists in the world. In fact, the worlds data is measured in exabyte’s (an exabyte is equivalent to 1M Terabytes) as of 2012. For more on the technical questions, check out the Gleanster Deep Dive report “A Crash Course on Big Data . . . for Marketers“.
It’s about Real-Time Analysis
Big data promises to give marketers the tools to conduct analysis on massive volumes of data, in real-time, across any filtering criteria. Want to look at billions of web interactions in aggregate and visualize them in some unique way – that’s where big data is heading. Want to optimize customer engagement through real-time business rules that personalize digital engagement based on every customer interaction over the lifetime (or potential value) of a customer- that’s where big data is heading. Solving the big data dilemma would allow organizations to capture, store, search, share, and analyze an aggregate data set, which in theory could provide new levels of insights for true data-driven analysis.
It’s about Structured and Unstructured Data
Structure information can be managed with a specific set of rules which uniquely classifies each piece of information within a database. A great example of structured information in marketing would be the fields on a form capture. Unstructured information on the other hand refers to information that does not have a pre-defined data model and it’s usually text-heavy information that could even include numbers, dates, and percentages. Twitter is a great example of unstructured data. The ability to analyze both structured and unstructured customer data at the same time could be a game changer for marketing optimization.
Linking Customer Data Across Fragmented Technologies is a Challenge – Even for Big Data
Multiple channels, means fragmented data. And that’s a problem because in the world of customer data, you need some kind of unique customer identifier to link customer interactions across channels. In theory, you would use an email address, a cookie, and or a registration process to start to link channels.
Roles will Change for Marketing
Access to big data means the expectations from marketers will likely start to rise. In fact, there’s a good chance marketing will have a new branch on the marketing organization chart called marketing scientists. These are marketers with heavy analytical skillsets that can transform big data analysis into strategy and actions that the average marketing leader can digest and translate into marketing strategy.
Marketers will Drive Big Data Investments
Marketers will likely be the driving force behind IT investments in big data solutions. Why? Because the business value derived from analyzing big customer data will lead to improvements in the top line. That’s spend that can be justified all the way up to the board of directors, shareholders, and CEO. Big data could help optimize marketing spend, tweak the marketing mix, improve products, segment customers, deliver real-time offers, improve the customer experience, increase revenue, and grow market share – all directives that keep the CMO up at night. So, don’t just dismiss big data as something for IT to worry about.
Moving Data: ETL Becomes ELT
There’s one more thing marketers should understand with respect to big data; data storage and data movement techniques have evolved, which ultimately makes big data a reality. Back in “the old days” (this is still common practice today) data warehouses were typically hosted on one or a few servers; servers that were limited in computing capacity. The real challenge was physically moving data from a variety of systems into a common data structure in a warehouse (think about the eclectic mix of marketing technologies that generate data – it has to be standardized for storage in a single data warehouse). Extract Transform Load (ETL) solutions became necessary because they could inject additional computing process prior to loading the data (because data warehouse servers didn’t typically have computing power necessary to Transform data). Information was routed from the source server, to an ETL solution sitting on a separate server, and then to the data warehouse. Unfortunately, ETL processes frequently run in batch modes at night or even weekly leaving a lag on information availability in the warehouse. Today, that process has shifted to Extract Load Transform (ELT) because computing capacity on the data server side makes it possible to do the Transform step after moving the data to big data hardware. (This is more technical than most marketers care about, but important nonetheless.)
_________________________________________________________
So there you have it. A quick and dirty “here’s what you need to know about big data.. for marketers”. Now you can impress your friends and maybe even your pals in IT. More than anything, this stuff is going to change a lot over the next 5 years because big data is still an emerging concept. Lots of different vendors will spin big data around product and service offerings and continue to convolute the definition. What matters is that you understand how big data may impact marketing in the future and why it’s not just a passing fad. Big data has the potential to unlock analysis on massive amounts of customer data. Data that’s currently doing little more than collecting dust.
For a more comprehensive look at Big Data for marketers, check out the complimentary Deep Dive report on Gleanster.com “A Crash Course on Big Data . . . for Marketers“.
