Several years ago, an Internet publishing company commissioned a detailed study to determine its audience. In the process, it uncovered a startling piece of data: bald-headed FOSS (Free and Open Source Software) users over the age of 40 represented a core part of its readership.
Since several of the company’s Websites focused exclusively on open source, nobody was surprised about the FOSS part. But bald users? Who could have expected that?
In the end, that data about follicle-challenged readers was good for little more than a giggle, because IT advertisers tend not to care whether a particular readership is bald or hairy as a pack of werewolves—so long as they have the power to sign purchase orders. And even though a couple of the company’s salesmen called advertising agencies that handled consumer products such as hair restoratives, they didn’t manage to make a single sale in that marketing space.
This is the problem with too much data: Beyond a certain point it becomes meaningless. It’s also subject to misinterpretation.
The “Abandoned Shopping Cart” Puzzler
Over and over, you hear talk from e-commerce merchants about the high rate of shopping cart abandonment. This is a phenomenon so heavily studied that virtually all merchants seem to know about it. But few realize the reasons people don’t continue their journey from shopping cart to purchase: high shipping prices, comparison-shopping or browsing, and more.
Anyone performing even the most rudimentary tracking of customer behavior on an e-commerce site can see that a whole bunch of people routinely click the “Add to Cart” button—and don’t follow through with a sale. The data is there. And yet many merchants refuse to dive into the reasons behind the abandonment.
Of course, that sort of detective work takes time. A merchant could always email customers to ask why they abandon e-shopping carts (or call them, if the Website collects phone numbers). They could also look at the Website and determine when customers first see how much is charged for shipping and handling. Many people add items to their carts, look at the shipping charges, and say, “No way!”
How do such essential bits of user data slip through the cracks to begin with? They end up lost in the Big Data pile that overwhelms everyone in an organization, from the CEO down to the marketing people. The Association for Information and Image Management (AIIM) discussed this very thing in a recent report (for which it demands a crazy amount of contact information). You can find a good summary of it at the Real User Monitoring Blog, but in the meantime it’s worth taking note of the report’s closing words:
“It’s a business problem and your job as an IT pro will be to guide your business units to the right solutions to find the answers they need inside that growing pool of data.”
Dealing with Big Data
Perhaps the best way to deal with Big Data—especially when you and your company first start to collect and analyze massive amounts of business data—is with the digital equivalent of an old-fashioned, bright red editor’s pencil. In other words, cut your stack of data down to a usable size by eliminating all information you are not 100 percent certain will help you cut costs, increase sales or otherwise directly and immediately impact your bottom line.
The AIIM survey cited cost and lack of in-house expertise as major Big Data utilization speed bumps. Cutting the amount of data you track can help smooth your way to more intelligent data analysis.
Once you’ve decided that 10 (or 20 or 50 or 100) points of information are the most important factors in how you relate to customers and suppliers, and even to your own employees, you can try altering prices, contact frequency, and other factors one or two at a time. In doing so, you might see those changes affecting your sales and profitability.
Or you may find that your customers really don’t care whether your packages are green or purple, so long as they’re easy to open. You might find that customers who consider a daily sales newsletter spam are happy to receive a weekly one with a week’s worth of bargains. And so on.
Once you’ve gotten your head around the idea that Big Data can easily overwhelm you, and cut your datasets to a manageable amount, you can begin using the data you’ve so painstakingly collected to help run your business more efficiently—and more profitably.