CHANCE: DOES IT HAVE A PLACE IN DATA?

By Shravya Kaparthi, Director, Analytics and Decision Sciences, RAPP Dallas

Marketing is all about influencing consumer choice by crafting a role for a product (or service) in prospective consumers’ lives. Oftentimes, this role comes by way of solving a problem, filling a gap, enhancing the current experience, or some mythical place in between — and then negotiating with consumer behavior to convince people of said product’s value and relevance.

A serious problem, though: There can be a lack of knowledge in how to articulate this value and create relevance.

For marketers to build strong brands around their product offerings, one important aspect at the core is investing time and effort into understanding the consumer. It’s all about the people, after all. People will inevitably define your brand, so it’s important to listen to them.

Fortunately, that’s what data is for.

Today, more than ever before, when the listening (and eventual understanding) is happening through the lens of data, we should never forget to decode the role of chance. Chance is, without a doubt, unavoidable in a person’s life, so the same should hold true about its role in data — at least with respect to consumer behavior in regard to marketing strategy.

And in order to understand consumer behavior, it’s becoming increasingly important to understand chance behavior. Only then can we truly demonstrate data-driven ingenuity as well as creativity in comprehending data.

The Likelihood of Chance
Despite all the sophisticated predictive analytics and attribution models available, there are lots of unexpected behaviors data cannot explain. Of course, statistics believes in the predictability of behavior. In fact, even chance behavior can develop a predictable pattern after a long enough period.

But this still doesn’t eliminate the need to interpret the role of chance in predictive calculations. Otherwise, it’s near impossible to leverage the insights necessary to power our creative teams beyond the obvious.

Take creative testing, for example. When reading the results, you must recognize that there’s only so much that data can tell us. Remember, “if you torture the data long enough, it will confess to anything.” Data makes sense only with good judgment.

Personally, I’m a data professional who believes in the magic of human behavior more than data-driven calculated numbers. If data backs it up, it makes us more comfortable, but any data point should not motivate or limit us from doing what we truly want to do.

When I read statistics like “90 percent of startups in the U.S. fail,” it clearly implies that failure is more probable than success. While still acknowledging the data point, powerful success stories would never happen if people used this statistic to stop launching new businesses. By that same token, would my aunt have shown a positive attitude and successfully fought her cancer if she had decided to believe the stats on cancer survivors?

To create powerful stories rooted in consumer truth, it’s imperative for marketing professionals to think through various aspects before getting to any insights or recommendations in regard to data. Here’s a quick guide to ask the right questions, when in doubt:

1. What is the source of the data?  Is the data reliable? Are there any biases in its selection or measurement? So much can happen from the time one plans a test or survey to when results are gathered. You want to know the data’s origin and whether it’s trustworthy.

2. What is the context of the data point? There’s always the potential of reading or interpreting data in a variety of ways. Consumer context can tell you a lot about a data point — and perhaps the role of chance, if any.

3. Are there any data anomalies? Errors in design and implementation of a test can lead to inconsistencies in data. Size and representation of a sample can also pose problems in tests and surveys. With bigger sample sizes, for instance, you’re less likely to get results that reflect randomness. That said, the greater the variation in the underlying population, the larger the sampling error can be.

4. Are there any outliers? When interpreting data, one of the variables often overlooked is the “outlier,” a data point that’s distant or different from all other observations. Again, this can be the result of an error in measurement or recording. But you still must question whether enough effort went into understanding this observation and the cause of it. Sometimes, it’s the outliers that give birth to new thinking.

5. Do you need to explore other aspects in the analysis? Constructive criticism is always healthy — this is where you question whether there are any reasons for the analysis to be misleading. Is there any statistical significance to a certain outcome? Knowing the statistical significance can help you quantify whether a result is simply chance or a factor of genuine interest.

If data is the most valuable new commodity, it’s important to believe in human truth around data interpretation. Data cannot create magic; the right interpretation of it does. Data professionals and creative folks must work together, bringing their guts to the numbers. They should marry the human truth of chance with statistics to enable informed, data-driven decisions.

Life is a probability, and it’s our stories that lie somewhere between the standard deviation of this likelihood and possibility. As marketing professionals, our duty is to understand what we can and can’t read from the available data to create powerful brand realities.