For marketers, it’s easy to feel under attack.

Virtually every new industry report emphasizes how much marketers spend on data analytics and how little they have to show for it. Every month, another expert jumps on the “Math Men” bandwagon, declaring that traditional marketing skills just don’t cut it in the new world of endless data.

This hyperbole is silly. 

Marketers face data challenges — no doubt — but the solution is not to treat data as some kind of oracle that replaces human decision-making. Despite the rise of social listening, predictive analytics and big data analytics, marketing will never be fully robotic. Data can tell marketers if revenues are going up or down, or it can give us a sense for what consumers are saying, but it can’t automate the intuition it takes to know which correlations and causes are worth exploring, nor the imagination it takes to dream up new ways to grow revenue.

Still, marketers’ data struggles remain a legitimate problem. The value of online consumer data has been self-evident for years, but IBM research indicates that as of last year, more than 80 percent of CMOs felt less equipped to deal with data than they did in 2011. What’s more, in a recent Deloitte survey of 300 CMOs, 71% said harnessing analytics is one of the biggest challenges they face. CMOs cannot over invest in “Math Men” hype, but they also can’t pretend traditional skills alone are enough to compete.

So if turning marketers into data scientists isn’t the answer, perhaps the answer lies in forming an alliance between human intuition and technology. Marketers haven’t struggled with data because they lack PhDs in Statistics — they’ve struggled with data because their technologies and techniques are incomplete.

Here are three tips to help marketers avoid the pitfalls:

1. Know the limits of popular techniques and ground yourself in revenue-based insights

Because incomplete technologies and techniques are part of the problem, the first step is to see where your current approaches fall short. For example, many social listening products track metrics like buzz, sentiment and virality. These metrics might be fine if you’re trying to increase user engagement, but if you need to ultimately justify your decisions in terms of revenue (and who doesn’t?), social listening alone doesn’t help. You have to ground yourself in revenue as a goal.

Popular techniques such as econometrics and predictive analytics are similarly problematic. The former tells you what happened but not why. The latter describes what is likely to happen, but not how you can change it. Each technique offers a few puzzle pieces — but not enough for marketers to construct the full strategic picture.

To succeed, marketers need to fill these gaps. Tie metrics to revenue, not to abstract concepts like “buzz.” Don’t focus on predicting futures, focus on shaping them. Don’t just look at when revenue moves, look at why it moves.

2. Understand how your data is cleaned

Recent research suggests CMOs are wasting millions of dollars on faulty analytics — that is, data sets that are tainted and inaccurate, and analytics methodologies that measure the wrong things. It’s no wonder over three-quarters of business decision-makers don’t trust data to support their business objectives, as Forrester recently pointed out.

These struggles point to an obvious problem: There is no pristine data set that provides a perfect picture into the world. Most data points have no explanatory value. The signal must be separate from the noise.

Suppose you’re the CMO of a beverage company. If your social media analytics tool flags tweets that say “This tastes great!” does it do so because “tastes great” seems positive and relevant, or because consumers’ use of the phrase correlates to revenue growth? Does your social data reflect your real customers’ demographics?

Without answers to these questions, marketers don’t have enough information to make informed decisions.

3. Be open to asking new questions

Analytics experts often advise marketers to approach data with specific questions in mind. On its face, this sounds like good advice. If you’ve just run a campaign targeting millennials, you probably want to find out how well you did with millennials.

But what if there were another audience that mattered even more than millennials? What if having just one specific question in mind limited the scope of what you could discover? In the “beverage company” example, it seems reasonable that “tastes great” is relevant — but maybe “taste” isn’t a strong differentiator for the brand. Perhaps consumers care more about how the beverage pairs with food, how smoothly it goes down, or whether it’s better served warm or cold. A CMO who focuses solely on “taste” would risk overlooking other, more resonant correlations to explore in the data.

Analysts have begun to call this concept “data discovery,” recommending that businesses invest in data products that augment human decisions instead of automating or replacing them. Finally, the industry is beginning to outgrow the myth that data can magically solve all problems, and moving toward a world in which data and human intuition work hand in hand.

Moves like this will help marketers avoid the pitfalls that have made analytics so challenging. Technologies and processes that integrate human intuition recognize the limitations of current techniques. Marketers need products that help them make better decisions, and the new class of offerings and techniques shows that established methods aren’t getting the job done. Just as importantly, these new methodologies and technologies demonstrate that marketers don’t need to be data scientists to succeed — they just need be clear on the rules of data engagement and how to align this with analytics.

Creative Commons Creative Commons Attribution 2.0 Generic License Title image by  stevendepolo