An analysis of your sales activity shows that 80 percent of your deals close on a Thursday. How does that affect the cadence of customer contacts you initiate early in the week? Or is it just a coincidence?
Questions like these are the type that data scientists tussle with every day, and they’re just the tip of the iceberg of correlation and causation questions that are emerging from large-scale data mining. Talk to Zeta Interactive Corp. CIO Jeffry Nimeroff (@JNimeroff) (right) for a while and you get the sense that we are still at square one in understanding big data’s potential.
Zeta Interactive’s database of information of 350 million people is just the starting point for understanding the behavior of consumers and business buyers. The company works with a network of information providers across a range of vertical markets to create rich customer profiles that enable marketers to better target promotions and guide conversations. Gathering data isn’t the challenge; it’s understanding what really matters versus what is coincidence or just noise – the correlation vs. causation conundrum.
Marketers don’t need to spend big bucks to identify patterns that can immediately improve their effectiveness, however. For example, the time lag between the arrival of an email message in prospects’ inboxes and their opening of that message indicates interest level, Nimeroff said. Ever more is revealed by tying clicks to opens and layering in time-series data. For example, it’s a safe bet that prospects who open a message within 10 minutes of its arrival and immediately click of an offer are going to be more receptive to a sales call that those who never open the message at all.
Much to learn
But in the age of big data there is much more to learn. Personal preferences gleaned from social networks can reveal, for example, that a prospect is a golfer. A sales rep can use that information to schedule a face-to-face meeting at a country club instead of a restaurant. “Even though I’m selling into a business, I’m dealing with an individual,” Nimeroff said.
That’s a particularly relevant point for B2B marketers. A 2013 study by Google and the CEB Marketing Council found that personal factors are nearly twice as influential in B2B buying decisions as in B2C choices. “B2B buying is extremely personal because purchases involve a variety of perceived personal risks – such as losing credibility, time or even a job,” the researchers wrote. “On average, the B2B customer
is significantly more emotionally connected to their vendor and service provider than a consumer.”
People’s use of digital media and devices is offering up a bounty of new information that can be used to better understand behavior. In the most recent development, Clear Channel Outdoor Americas yesterday said it’s partnering with several vendors to track people’s behaviors by tapping into their mobile phones. The data will be used to customize billboard advertising to the demographics of people who might be passing by.
“If I have 200 apps on my phone, that is a window into what I like,” Nimeroff said. “It may paint a picture of me as a sports guy who travels a lot. That gives marketers the ability to compile repositories of data about me.”
Keeping it anonymous
If all this sounds a little creepy, take comfort in the fact that laws are pretty strict about how specifically personal information can be targeted. In most cases, data is aggregated and anonymized to create profiles. Cable television providers, for example, can collect statistics about the individual households they serve but can only sell that data on a rolled-up basis.
So much new data is coming on line that that data scientists won’t run out of things to do anytime soon. The Internet of Things will offer up a bounty of new statistics to correlate. “My heart rate monitor collects a sample per second. This rich and temporal data can be combined with other rich and temporal data, such as email responses,” Nimeroff said. “When we start to align all of these data sources, very powerful correlations can emerge.” For example, linking heart rate data to TV viewing activity can identify programming that evokes an emotional response.
So how does that relate to the question about why all those sales close on Thursday? Nimeroff said data science may never provide a definitive answer. What’s more important to understand is that it’s one data point that can be useful in understanding the buying process. Marketers need to understand that a sale is a series of conversations that takes place across multiple channels over time. “You need to think about how you’re going to tailor that conversation,” he said. “The independent channel strategy has to break down and channels need to work together to enable a more holistic approach to marketing.”
Ultimately, analytics can guide decision-making, but gut instinct still matters, he said. “Data science practices can support intuition.”