This article was published as a guest post in AllThingsD, and is republished here for Digital Quarters readers.
The New York Times named 2012 the crossover year for Big Data: as a term and as a concept, Big Data broke through from the tech circle and into mainstream consciousness. (So much so that even Dilbert’s boss was talking about it.)
We’ve seen huge advances in our ability to generate, collect, and store an explosion of data points: 90% of the world’s data has been accumulated in the last two years alone. We’re generating 2.5 quintillion bytes of data daily, and every serious company is dutifully logging and contextualizing every impression, every click, and every purchase with excruciating detail.
That said, shockingly little happens to the information once it’s been stowed in the database. A good friend gave voice to this dirty little industry secret the other day:
“Nobody wants to use the data.”
He’s remarkably spot-on. Even though almost every CEO says their company is becoming data-driven, the fact is that most high-level decisions are still being made from bullet points, not data points.
What the data revolution brought us was systems for collecting data – but collecting is the easy part. And even more importantly, it’s the safe part.
The Real Problem: Data Phobia
The trouble with data is that it asks as many questions as it answers. Your engagement is down, bounce rate is up, search traffic is up – why is that, and what can we do to make it higher, lower, and higher? Data almost never hands you the answers or insights directly; it just illuminates the issue. And it illuminates a whole bunch of them at once, so it’s up to you to figure out what the priorities are.
If this problem is an “opportunity in disguise,” most executives seem quickly scared off by the masquerade. In truth, Big Data raises the bar for how smart you have to be as an executive.
The easy answer – leaving the analytics to the analytics department – relieves you of the responsibility of figuring it all out, as though it’s unknowable to anyone without a degree in data science. But it also relieves you of the answers.
What is the executive’s greatest fear? That exposing the trove of data without knowing what to do with it makes them look worse, not better. In media, many have hidden that fear behind the veneer of idealistic purism. I remember talking with Martin Nisenholtz several years ago when he was at The New York Times about how data is used in a newsroom; I asked what would happen if he shared performance metrics with reporters in real time (obviously this was before Chartbeat) to see what their audience cares about. He said, “They would throw me out.”
Our strong institutions and professional commitment to standards have ensured the journalistic values of truthfulness, accuracy, objectivity, impartiality, fairness, and public accountability. None of those values are furthered by closing our eyes and ears to our own audiences. The result is a paradoxical culture that boldly states “content is king” and yet refuses to quantify its value for fear of tainting the purity of the product.
The Opportunity: Using Big Data to Make Big Bets
Until recently, we have had startlingly few case studies of the transformative power of Big Data on which to model our own big changes in media. Instead we’ve had IT initiatives that promised big insights, but ended up delivering big databases and bigger IT bills. For once, it’s not the IT department’s fault – it’s those of us who are using the data (and, more often, aren’t using it) who are to blame.
That’s why I turn to those who have made the big bets to see what’s different. Netflix has long been the poster child for using data to drive results, and now they’ve proven in no uncertain terms that when you ask your data the right questions you can find hugely valuable insights – even in the sacred domain of content creation.
Before Netflix pursued the option to buy House of Cards, they looked to their massive data stash. They wanted to know: Do Netflix users enjoy political thrillers? Check. Of political thriller enthusiasts, how many also watch David Fincher films? A whole bunch. Oh, and one more thing: Is this crowd fond of Kevin Spacey? As it turned out, there was a very healthy crossover in that Venn diagram.
Not only did this insight give Reed Hastings the confidence to bid on House of Cards – it gave him the level of certainty necessary to outbid heavyweights like HBO and AMC for the series.
What I love most about this story is that the questions were so simple, so logical. Sometimes the sheer volume of data at our fingertips overwhelms us and makes us forget that the fundamental strategic questions haven’t changed. What has changed is that now we have far better access to the answers. And when you can give your users what they want based on the signals they themselves have been sending you, that’s when Big Data starts to earn its keep.
Five Questions You Should Be Asking Your Data
Forget about Omniture and Google Analytics and all of the data minutiae you’re already tracking. Forget about little personalization features. The most valuable data doesn’t fit on the dashboard. Think bigger and move upstream: what’s the most amazing new product or service you can create? Here are five places to start digging:
1. What does my audience LOVE? Cut the data every way you can to deeply understand this, with nuance – then reorient around that product. It might be parenting advice, or current memes, or breaking news. If you can find a common thematic thread in your most-consumed content, you have a great starting point for further segmentation. Lauren Zalaznick turned the Bravo network around by pinpointing the five key interests of the audience, cutting out the clutter, and giving them more and more and more of what they loved (hence the hugely popular Top Chef and Real Housewives).
2. How do they want it? Netflix noticed that a significant number of users were watching marathon-style, and so they bucked TV tradition and released House of Cards all at once. How could you change your content packaging to better match the real habits of your users? Many have tried and failed with full-length video programming on the web; that’s because (so far at least) most Internet audiences can’t sit still long enough to watch a 30- or 60-minute program. Adapt your delivery to what your audience wants.
3. How can I best relate to them? Personality is critical – so which of your brands’ public talents and personalities relate to whom? It might be a popular columnist, Don Draper, or Boo the Pomeranian. Figure out which personalities your audience connects with the most, and leverage them into the other themes and packages.
4. What secret signals is my audience sending? Target famously figured out how to identify pregnant shoppers and even estimate their due dates months before the woman ever purchased a stroller or a pack of diapers. Find out which clues in your data indicate that a customer may be on the path to a new phase of life, and start messaging them with your relevant content even before they get there.
5. Where is my sweet spot? Once you discover the key themes, packages, and personalities that resonate with your audience the most (and at which relevant life stages), you can cross the data sets and identify your best untapped opportunities. Don’t just tweak your existing products and advertising – create whole new products that are designed specifically to thrive at the intersection. Just as the strong affinity overlap for Spacey / Fincher / Cards gave Netflix the confidence to make a bold bet, your own Venn diagram will spotlight your best chances to create knockout content that is destined to succeed.
Rethinking Management: Ask, Understand, Execute
When it comes to dealing with Big Data, our skills haven’t evolved as fast as our capacity. We all have a functional specialty, whether it be content creation or distribution or sales or management – so whose job is it to ask the right questions of the data? Big insights and actions aren’t led by a data scientist; they are led by an executive who has an integrated view of customers, products, distribution, and sales.
But asking Big Data the right questions isn’t just a new practice to add to the management to-do list. Pulling it off requires a rethinking of the manager’s role entirely. We’ve traditionally thought of management as the discipline of managing people and managing the business. Now it’s time to add “managing our understanding” to the job description.
The time of the executives who merely “execute” is past. The successful executives in this post-Big-Data world first ask, understand, and then execute with the full support of the data behind them.