Skip to content

How To Find the Truth in Data and Marketing Analytics

How To Find the Truth in Data and Marketing Analytics

What are your brand’s “left ventricles” waiting for discovery?

Seth Stephens-Davidowitz posed a variation of that question in his keynote presentation at the Marketing Analytics & Data Science Conference.

Seth Stephens-Davidowitz is standing on Content Marketing World 2024 stage wearing glasses and a grey suit with a pink button down.

Let me explain.

Seth, a data scientist and New York Times bestselling author, recounts the story of Jeff Seder, an entrepreneur and self-described innovator in analytics for horse racing.

With three degrees from Harvard University, Jeff quit his banking job to be around nature and horses. His mission was to use data to determine what makes racehorses great. It was an era before Michael Lewis’ 2003 book, Moneyball, explained the use of analytics in Major League Baseball.

How do you find the winning data analytics formula?

Jeff analyzed the attributes of racehorses to see which predicted champion-level success. The characteristics he examined included:

  • Size of nostrils
  • Volume of fast twitch muscles
  • Size of defecations

They all failed to predict success. But Jeff kept trying, and after 20 years, was on the verge of bankruptcy.

But then the tide turned.

Jeff built the first electrocardiogram to measure a horse’s internal organs and discovered that the heart’s left ventricle was a valid predictor of champions. Case in point: American Pharoah. In 2015, the horse won the Triple Crown, the first in 37 years to win all three famed races run by 3-year-old thoroughbreds. But two years earlier, no one realized the amazing racehorse American Pharoah would become, except Jeff.

Consider these attributes of American Pharoah:

  • Height (56th percentile)
  • Weight (61st percentile)
  • Pedigree (70th percentile)
  • Size of the left ventricle (99.61st percentile)

While height, weight, and pedigree weren’t overwhelmingly impressive, American Pharoah’s outsized left ventricle accurately predicted its Triple Crown success.

Seth shares the story to illuminate these takeaways for data-focused marketers:

  • A dataset’s value usually is not its size but its newness.
  • Winners in Big Data are entrepreneurial.
  • Winners in Big Data fail a lot to find the big winner.
  • “Left-ventricle” discoveries are out there.

By embracing these insights, Seth says, you can make your data model 10 times better than everyone else’s.

MADS lesson: You can discover your brand’s “left ventricles” in analytics data (e.g., web, email marketing, social media marketing, paid search). Form a hypothesis by asking a question. Analyze the data to answer it. If the answer is no, ask another question. Keep going, like Jeff Seder did, until you find the “left ventricle” that leads your campaigns to Triple Crown success.

Can you trust what people say?

All data, though, isn’t necessarily helpful. Take surveys, for example. Seth explains that the format frequently used by Gallup and Pew Research Center can be used to understand why people do things. But there’s an inherent problem with surveying.

People often say what they think will impress someone else. It’s called social desirability bias. Seth gives an example of people saying they voted when they didn’t or saying they voted for one candidate when they voted for another. They knowingly give incorrect responses because they want to be seen as doing the socially acceptable thing.

But where can marketers go to find pure honesty? In Google searches, Seth says. He calls it the “digital truth serum” because people confess to it. They ask questions about their health conditions, relationships, and whatever else is on their mind. They feel free to be honest because they perceive no one is on the other end to judge them.

Seth says Google knows more about people than their own partners and family members. So, instead of asking what people do, use Google Trends. It is better than Gallup at predicting who will turn out to vote and unemployment rates, and at measuring racism.

However, Seth cautions us not to treat all Big Data as the same truth serum. If Google is a digital truth serum, Seth calls social platforms like Facebook the “digital brag-to-my-friends-about-how-good-my-life-is serum.”

To illustrate his point, Seth shares his research about the term “husband” in social media posts and Google searches. Here are the top phrases completing the statement, “My husband is ___________,” based on the platform:

Social mediaGoogle searches
The bestGay
My best friendA jerk
AmazingAmazing
The greatestAnnoying
So cuteMean

As you can see, spouses want to project positivity about their husbands on social media, whereas Google searches reveal what they really think.

MADS lesson: Study what users say (i.e., survey data) and what users do (i.e., web or app analytics data). Find your source of digital truth serum and use that to guide your marketing and business strategy decisions.

Should you believe the gut?

Seth says everybody lies. He even wrote the book Everybody Lies: Big Data, New Data, and What the Internet Can Tell Us About Who We Really Are.

It’s not just in surveys and social media posts. People lie to themselves. In other words, your gut feelings can be wrong. Seth tells a story about an app called Mappiness. Users receive a ping at different times during the day with a prompt asking how they are feeling and what they are doing.

The first-person data revealed that these activities are most associated with happiness:

  • Intimacy/making love
  • Theater/dance/concert
  • Exhibiting/museum/library

These activities are the least associated with happiness:

  • Care or help for adults
  • Working/studying
  • Sick in bed

While these two lists may not surprise you, the research revealed a surprising result — the difference between what people think will make them happy and what actually does. Here’s what the Mappiness data uncovered.

Underrated activities that made people happier than expected included:

  • Museum
  • Sports
  • Drinking alcohol
  • Gardening
  • Shopping

Overrated activities, which made people less happy than expected, included:

  • Sleeping
  • Computer games
  • Watching TV
  • Eating
  • Browsing the internet

Seth says the data is clear: If you want to be happier, spend less time inside. Go out, be active, and explore the world.

MADS lesson: Collect data to analyze how well your marketing gut (i.e., perception) performs in reality (i.e., overrated, underrated, on target).

What role does appearance play?

Seth brought the personal into the conversation when citing research from Alexander Todorov, a professor at the University of Chicago Booth School of Business. His study explored the prediction of election winners based on their appearance.

In the experiment, participants viewed two candidates’ photos side by side and were asked to identify which one looked more competent. The research concluded by finding that 70% of elections were won by the candidate whom users selected as appearing more competent.

Before reading this research, Seth says he never paid attention to his appearance. But now he considered changing it, so like any good data scientist would, he turned to data analytics, for the assist.

Using AI-powered FaceApp, Seth created 100 versions of himself. Following Alexander’s research, he placed two versions before people and asked them which one looked more competent.

Using a regression analysis, he concluded that the only two factors that influenced the competency preference were a beard and eyeglasses. Wearing both boosted his competency rating to a 7.8 from a 5.8.

Not surprisingly, Seth appeared on stage with a beard and glasses.

MADS lesson: Just as a candidate’s appearance influences the viewer’s perception of competence, so do the look and feel of websites and apps. Use focus groups to understand which design elements users associate with feelings of delight or satisfaction.

How will you use data analytics in your marketing?

I left Seth’s keynote talk inspired to find new and creative ways to use marketing analytics to drive decision-making. Now, though, I will do that better because I have heard Seth’s “left-ventricle” thinking. Data should be the arbiter of decisions, but only if it’s true and accurate.

All tools mentioned in this article were suggested by the author. If you’d like to suggest a tool, share the article on social media with a comment.

Want more insight from these and other Content Marketing World speakers? Register for an on-demand pass to get access to session recordings through January 31, 2025. Use code BLOG100 to save $100.

HANDPICKED RELATED CONTENT:

Cover image by Joseph Kalinowski/Content Marketing Institute