Data Vs. Intuition: How To Approach Data for Great Marketing

Marketing data analysis tips for the savvy marketer

Share on FacebookTweet about this on TwitterShare on LinkedIn

When big data first started to become the next big thing, the conversation was about holding human judgement up against the numbers to see who’d win: the robots or us.

We’ve come a long way since then. Most modern marketers are well aware that today’s data/intuition dialog doesn’t have to be an either/or debate. It’s an and/also conversation. Data and creative problem solving are symbiotic, and in the current world of digital relationship marketing, one is rarely helpful without full use of the other.

Carey Nadeau is the Founder and CEO of Open Data Nation, which is a social benefit company that combines open, public data with data science techniques to increase transparency and productivity of public agencies. We asked her to weigh in on the data/intuition dance. Nadeau’s work seeks to use publicly available data to predict outcomes, help prioritize resources, and establish data-driven performance metrics. In her own words, “Open Data Nation builds predictive statistical models to say, based on the data from yesterday, where do we anticipate the problem will be tomorrow?” And this task—of building a crystal ball out of data and anticipation, and passing those numbers and ideas through an understanding of the world you’re dealing with—is not all that different from what’s asked of digital marketers every day.

We can apply Nadeau’s frameworks for working with data to the task of understanding what makes a user or customer tick (or spend, or download, or whatever action will drive them further through the funnel).

Marketing and predictive analytics

The mobile marketer’s task, boiled down to basics, is to predict what the user or customer will do next, in order to position themselves in such a way as to be there, waiting for those anticipated user behaviors to happen, with a marketing strategy in hand.

To accomplish this, it makes sense for marketers to hone in on characteristics and user behaviors that correlate, or seem to predict certain outcomes. If you’re collecting the data that make the most sense for your task, this is the first step in predictive analytics. Many marketing automation systems today are enabled with some predictive capabilities. Perhaps the most basic example is running and A/B test, showing two smaller segments different versions of a campaign, and then sending the best-performing variant to the rest of the user base. You’ve predicted what will perform better based on the data you have.

But this doesn’t mean qualitative analysis is out the window. When we dig into our user data to look for patterns that will enable our predictive analytics tools, we must first “develop a qualitative understanding of conditions,” says Nadeau. In less mathematical terms, this means using regular old human judgement to look at a scenario and parse ideas using only the power of the humble human brain.

Develop a theory. Then test it.

In short, we have to start with an educated guess—a guess based on what we know from our own human experience, or based on what we’ve gleaned from prior research, prior experience in our jobs, and observations as marketers and as people being marketed to—in order to come up with a theory. Then you test that hypothesis by testing your own data, or by exploring existing data, to see if it’s already been proven.

So while data is at the core of the processes Nadeau describes, and that she in fact has built a career on, our own intuition and inklings must plant the seeds in each data analysis.

Facing the challenge of unwieldy data

When jumping in to begin your analysis, Nadeau says the most surefire way to keep it simple and cut right to the chase is to hold your thesis close at hand. “Keep in mind the purpose for exploring the data.” It will be your flotation device in that data deep-end.

“Analysis paralysis” can make for a frustrating deadlock. Bringing in some common sense, creativity, and intuition can help.

You can “cut the data nine different times in nine different ways,” says Nadeau, but it doesn’t help you or your team or your customer to present or act on every possible iteration at once. “You go explore a hypothesis. Sometimes you succeed. Sometimes you fail. Sometimes you find new questions to ask. At the end of the day,” Nadeau says, “you’ll cut out 90% of the process, and focus in on the 10% that helps you tell your story.” This means that a lot of the legwork you do might not end up pointing to the question you’re trying to answer after all. That’s ok. Keep your purpose in mind and keep looking.

There’s nothing wrong with saying, “We thought this would be important, and we ultimately found out it didn’t contribute to the story we wanted to tell.”

When your theory is wrong

It’s important to not set out to prove your theory, but to discover whether in fact it’s true or not. “In data science, sometimes you don’t prove what you set out to prove,” says Nadeau. “Say you have a theory, and your analysis shows the exact opposite. Your theory is dead wrong. It has a negative effect on the outcome. To say otherwise would be dishonest.”

In short: have a hypothesis, but don’t be overly invested in your own narrative. You have to start with a hypothesis, but even when it’s wrong, there’s still something interesting to be found there.

The questions to focus on, she suggests, are, “With what you have, can you affect change? Can you move the needle?” Can you make good business decisions with these perhaps surprising results?

Avoid making false connections

“When we prove something with statistics,” says Nadeau, “we know that correlation exists, but we don’t know if they’re causal. It’s a core fundamental of statistics.”

Just take this hilarious website that creates graphs based on spurious correlations. From 1999-2009, for example, there was a clear statistical correlation between the number of people who drowned in swimming pools as compared to the number of films Nicolas Cage appeared in. These two data points are clearly unrelated—common sense can tell us that one thing is not causing another—and yet the data makes a clear case for the relationship.

“We have wrong theories about the world because we’ve proven them with statistics,” says Nadeau. This could just as easily happen in your role as a marketer as anywhere else. “There’s always an opportunity for theories we thought were true to stop being true.” Likewise, things we thought were true sometimes just aren’t true.

There are caveats to every analysis

It’s easy to get hung up on all the different ways one can look at the data, or second (and third and fourth) guess what you might have missed. Were there more observations to be gleaned?

If that doubt is there, suggests Nadeau, acknowledge it. Mention it to your team or your boss, or whoever’s helping you make decisions. Discuss these caveats, but do so in a way that presents the solution first. Focus on the tangible.

There are questions that data alone cannot answer

Data is here to stay, and at this point, even the most data-driven among us has likely experienced the acrobatics of finding a useful middle ground between the numbers and what simply feels right. Even a die-hard number cruncher like Carey Nadeau.

Data can be informative as well as instructive. It can tell a marketer when to target a campaign, who to focus on in a given moment, and which themes or contexts will be most effective. It often can’t tell you, however, things like how important your findings are, held up against outcomes that have occurred in the past. Or how correlated a variable is to a certain outcome.

There are other considerations that data can influence, but can’t dictate, like how precisely to interpret findings in such a way as to create a cohesive and compelling ad or campaign. That noble task is left to a vital, powerful machine: your brain (with a dash of heart thrown in, of course). Check out our companion post: Data Vs. Intuition: Why Your Marketing Needs That Human Touch.

Share this post

Share on FacebookTweet about this on TwitterShare on LinkedIn