by Bruce Nickson
It’s not about the DATA, it’s about the BIG.
I thought it might be an idea, in advance of the April 17th Breakfast Speaker Series (BSS) event, “Utilizing Big Data to Your Advantage”, to find out exactly how big Big Data is. Well it’s bigbig. Something along the magnitude of 709 exabytes of information flows through the Internet annually. And it’s growing fast.
What the heck is an exabyte? How big is it? It’s huge: just 5 exabytes would be a transcript of all words in all languages ever spoken since the beginning of human language. Or, to send an exabyte of data using 747s, you would need to fill up 13,513 747s with 4.7-gigabyte DVDs. Measured end to end, those 747s would stretch for 640.25 miles (1,031 kilometres).
During her introduction to the BSS event, Sarah Clayton, the President of the BCAMA and Marketing Director of Regent College, described her own experience with the data held by Regent, including the social statistics, the web analytics, the organization’s CRM and more. She felt she needed a mathematician to make sense of it all, with not a mathematician in sight. An experience, I think, typical of many.
Richard Muller – Principal, Sum Things Ventured
The first analogy presented by Richard Muller, the moderator, came from Geiko’s CMO Ted Ward, when he described managing the volumes of data coming through the cloud: “… as for the cloud, I leave it to our IT folks to juggle the Jell-O…”
The data is not quite solid, but solid enough to use. What Ward is referring to as data is a management issue.
So, for all marketers, the question becomes: is Big Data an obstacle or an opportunity? Apparently, the majority of marketers see Big Data as an obstacle in one form or another.
The Holy Grail of data manipulation is predictive analytics. Which means, quite simply, can we get the jump on our competition? Can we engage our customers in more meaningful profitable ways through Big Data analysis? Can we see around the corner into the future before the others do?
BIG ≠ Profit
Jonathan Flander – Vice President, Global Data Strategy, Wunderman
The presentation from our first panellist, Jonathan Flander, might easily have been called “Harnessing BIG for Profit” using emotional and connected data. Or “fishing where the fish are…”, which is a cryptic way of saying finding where the consumer is providing meaningful sources of data. Which in turn means understanding cross-web behaviour? Got that? (I’m not sure I do.)
Jonathan went on to say “it’s about the outcomes”. And further about slicing the segmentation pie even finer to provide insight that can be acted upon tactically, as in designing a new campaign, as well as strategically, as in managing an overall data capture and management platform.
The contrast from a couple of years ago (2008) versus now and in the future allows us to see complexities that we weren’t previously aware of. This is due largely because of the onset of digital, mobile and social. Plus, the consumer is always interacting with your brand – 24/7.
Using the traditional method of collecting data is not working any more. You can’t get to where you want to be from there. So, big data management is about reaching out and finding the right data sources, and about developing a framework that prepares you for the fact that next year there is going to be a new data source, a new Pinterest or a new Instagram. Marketers need to understand where consumers are engaging with brands, and where they are interacting with them.
While traditional data sources will allow us to understand transactional behaviour (e.g., “22% of her purchases are health and wellness related”) as well as basic demographics, emotional data allows for a more complete holistic and connected relationship with your target.
The correct use of Big Data allows marketers to discover how the audience is engaging with us, how they are searching for us, and what they are saying about us. It further allows marketers to determine who they should be spending time with, and developing messaging and conversations appropriately.
The marketing organization must begin to think about the management of data. The data acquisition platform must be established. This begins by mapping out your customers’ journey, and how customers interact with your brand. It doesn’t mean listening only to sales or customer service, but collecting all of the data from the various silos and connecting them to see the customer journey map. Then you might be able connect the journey to external data sources. And when that happens, you’ll have meaningful actionable information.
Pushing Data into a Warehouse ≠ Business Insight
Jesse Gross – Senior Manager, Ernst and Young Digital Analytics
Our second panellist, Jesse Gross, began talking about numbers, specifically “cardinality”. At this point, I had to Wiki the concept of cardinal. Whew… And then I thought “For once, couldn’t data guys speak English?”. But then I remembered that one of the best digital agencies in Vancouver is Cardinal Path. They speak English.
There is only so much that the human brain can endure before the rapidity of the information overwhelms our ability to process. So here is my impressionistic take on Jesse’s talk.
The collection of data for data’s’ sake does not yield results.
None of this addresses key questions such as: • Who are my best customers? • How can I increase awareness? • What are the key variables to improve site performance?
Jesse suggests segmenting your data into tiers, of which he suggests two; tier 1 is called “who”, and tier 2 is called “why”. Get a good grasp of your metrics. Develop signatures checked from the who and the why. This will flow to checking projects. The two-tier segmentation model is applied to all data collection conduits, including survey data, mobile app use, social activity across all pipes and venues, web visits and so on.
So survey data across pipes might begin to look a bit like this:
The problem isn’t not enough data, but too much. Especially if you layer in all other data pipelines and sources (which is coming at you 24/7), and the variables that they present from a demographic and other metrics point of view. Multiple pipelines with multiple metrics streaming at you as fast as an electron can move. That’s fast.
So what is the solution to all of this? Well, you need a warehouse. That’s for sure. Doesn’t much matter what form it takes – Jesse is pretty agnostic on this issue. But on the GIGO principle, you’re going to have to figure out a way to sculpt this data into a form that offers insight into customer behaviour, competitive reactions, market tendencies and all the good stuff that marketers are supposed to be good at (and, too frequently, aren’t).
The advice in the end, is not a great “ah ha” moment. It’s closer to getting back to basics, albeit with a layering-in of serious technical savvy. The population must be segmented. The variables must be analyzed. The metrics must be counted, sliced and diced. And since you have the data drawn from all sources at your disposal, you are now ready to start asking questions about your targeted universe that have not been answered.
And then you create a model that represents all the known information you have or can gather about the universe under consideration. It seems likely that you’re still going to come up with a “persona” analysis, but with richer, deeper and faster information with which to create that persona. That persona might actually begin to resemble someone you know: i.e., a real person in real time.
In the End …
Let’s just say that space constraints disallow me from reporting on the Q&A that followed. I can say with accuracy, I think, that some people were puzzled, skeptical and possibly confused, but everyone was definitely more enlightened as to current thoughts and directions on the Big Big.
There is no question that Big Data is the buzz. It has not just infiltrated into the marketing and digital space, but also pretty much every industry you can think of. Practitioners of all disciplines have seen the great graphics, and the wonderful video. And many are wondering if this means marketing is turning into an “unleash the bots”, rather like conducting a drone war with proxies. The more information your bots collect, the more you rescue the damsel, triumph over the dark knight, and then look around the corner to a bigger cache of gold.
Bruce Nickson (@brucenix) works in marketing and sales of engineering services. He once held the job title of Executive Director.
Editor’s Note: Many thanks to our moderator and panellists for their presentation, and for allowing us to reproduce some of their slides above.