SPEAK VISUALLY Receive practical tips on how to
communicate visually, right in your inbox.
You may not know it, but big data is the next big thing in content marketing.
Contrary to the popular belief that data-driven content marketing is a field reserved for data scientists and statisticians, the creation of this type of original and insightful content is the next logical step in an online world where almost everyone is now a content creator.
To see how little this unique tool has been exploited, just Google the term “data-driven content” and you’ll mostly see references to how to use audience data and insights for creating targeted marketing strategies and campaigns.
The problem with this up-and-coming trend, however, is that if you look at all the companies who have ventured into this somewhat new territory, you’ll find a few companies that are doing it well--and many that are not.
When dealing with information in the form of data visualizations and infographics, there’s always the risk of misinforming audiences by misapplying proxies (which are substitutes for a particular statistic or indicator), assuming that correlation means causation or oversimplifying data and forcing it into a binary scheme (for example, blue states versus red states).
To help you in your journey to becoming a data-driven content creator, we’ve compiled a list of brands that are doing it right--and some that are not.
As expected, one of the pioneers in producing branded data-driven content is Google. Given the huge amount of proprietary data the giant has at its disposal, it’s no wonder it’s one of the key brands making headway into this largely uncharted territory.
In this interactive piece (click to view full-screen version), the most popular music genres among Google Play Music users are visualized using a Music Timeline. The thickness of each stripe represents the relative popularity of a particular genre during a specific year.
This fun piece combines tons of data in an interactive visualization that can be explored for hours on end. Simply click on each stripe to view increasingly specialized genres, each with its own breakdown by artist and/or album.
Another brand that is rocking data-driven content creation is Jawbone, a wearable products company based in San Francisco.
Like Google, it used proprietary data collected by its UP fitness tracker systems to map the movements and sleeping patterns of users in seven different cities--New York, Paris, Beijing, Moscow, Dubai, Tokyo and Madrid.
Among other insights, the interactive piece reveals that New Yorkers are the first to go to bed and also the first to rise, while users in Tokyo go to sleep later than those in other cities but wake up earlier as well. The data also revealed that users take afternoon naps in Beijing and Madrid and up to 10% of users sleep in until 10 am in Moscow and Dubai.
The British price comparison website Gocompare created this colorful visualization of migration data, which registers moves between regions in England and Wales.
Users can choose different motives for moving (such as work, retirement or going to school), as well as the age range of movers and different regions on the map.
By representing each individual as a single colored bubble, the viewer can come away with a comprehensive bird’s eye view of the migration patterns that marked the United Kingdom in 2014.
This cloud-based healthcare tech company created this interesting interactive map, which visualizes health search patterns among U.S. employees who use its platform.
Some of the most interesting findings include the fact that men and women used different language to search for their healthcare needs. Men, for example, tended to use natural language to search for a specific problem, such as “rash” or “tooth,” while women used doctors’ terminology, such as “colposcopy” and “endometrial ablation.”
Additionally, it revealed generational differences in search patterns. Millennial employees, for example, were more likely to search for dental care and anxiety disorders. Also, users in Southern regions are three times as likely to search for weight-loss surgery and obesity-related issues.
Another brand making a name for itself in the world of data-driven content creation is General Electric. A multinational conglomerate that makes a wide variety of products and collects data on just about everything, from healthcare to energy consumption, GE has bet heavily on data-driven content creation--even dedicating an entire section of its site to data visualizations.
For example, in the mesmerizing video seen above, GE visualized the amount of power output registered by 713 of its giant turbines found all over the world. Each blue line represents and gas-powered turbine and the height of each represents its capacity. Additionally, inside each sliver, there is a line graph that shows the exact power output over the course of a specific day.
For music trendsetters the world over, there’s this fun Web app created by the music streaming service Spotify. Using your account details and its own data on breakout artists, the tool, appropriately called “Found Them First,” determines whether you were one of the early listeners of a now-popular band or artist.
Spotify considers musicians with more than 20 million streams and a growth rate of 2,000% between 2013 and 2015 to be “breakout artists,” while only users in the top 1 to 15 percent are considered “early listeners.”
This go-to resource for homebuyers, renters and sellers has been creating data-driven content for years.
Take, for example, its Crime Maps tool, which allows users to explore and compare crime rates across the U.S, which is intended to help consumers make informed decisions about where they want to live.
Using geodata from partners such as CrimeReports.com and EveryBlock.com, it allows users to view crime rates in a specific area, view different neighborhoods at once and compare crime statistics of two different areas. Users can also provide comments and feedback via Facebook.
This Russian antivirus firm launched this impressive interactive cyberthreat map, which visualizes all the cyber security incidents occurring around the globe in real time.
Using information collected by the Kaspersky Security Network, the map summarizes the data sent by protected devices and visualizes them by using different colors to mark different threats. Users can rotate and zoom into any part of the globe, as well as view a more detailed description of each threat and share it on different social networking sites.
While it may be easy to let all the beautiful colors and effects impress us, it’s also worth mentioning that data-driven marketing has its pitfalls. It can be very easy to make amatuer mistakes and, in the process, unintentionally disseminate inaccurate information.
Statistician Jake Porway, for example, argues that human beings tend to rationalize beliefs with data--and not vice versa. Since we often don’t know how the data was collected or manipulated, we should always view data visualizations with a healthy dose of skepticism, as we do with any other visual or textual content form.
Here are a few examples we can learn from:
Jacob Harris, senior software architect of the New York TImes, explains why Pornhub’s study on who watches more porn, Republicans or Democrats, was seriously flawed.
He observes that one of the most noteworthy anomalies in this study is that Kansas, a red state, exhibited an abnormally high amount of porn consumption per capita.
He explains that this is the result of a process called geocoding, which often produces less than accurate measurements. When a large amount of IP addresses cannot be attributed to a location more specific than the U.S., then the point returned by geocoding is a point in the center of the country’s map, which is, of course, Kansas!
Another example of data-driven content gone awry is this Sexual Wellbeing Survey conducted by Durex. According to this poll, people in Nigeria and Mexico have the most exciting sex lives, while the French report some of the lowest levels of sexual excitement.
Clearly intended for attaining virality rather than informing, the study neglected to take into account various variables that could potentially skew the final results.
For example, Nigerians were interviewed in person while the rest of the participants completed the survey online. This is certainly cause to bring out the skeptic in each of us. Would you respond in the same manner if you were talking to a real, live person about your love life rather than a computer?
If you’re in the mood for more mind-numbingly bad examples of data-driven content, here’s another one by the dating site OK Cupid.
In its piece 10 Charts About Sex, it erroneously assumes that correlation means causation. After mapping per capita GDP and finding a close correlation with the portion of people looking for casual sex, it mistakenly concludes that “money seems to be a more powerful influence on sex drive than culture or even religion.”
In a past post on the best infographics of 2015, we showcased a series of humorous graphs intended to prove the point that correlation most certainly does not equal causation. For example, did you know that there’s a strong correlation between the divorce rate in Maine and per capita consumption of margarine?
In the introduction, we mentioned misapplied proxies as another huge mistake to avoid, as exemplified in this piece by the real estate company Movoto.
Apparently, from the methodology outlined by the folks at Movoto, it seems that stress levels can be accurately calculated by simply adding all these factors together:
So, the reasoning goes that if you have a lower median household income, then your stress level should be higher. Or that if your neighborhood has a higher percentage of uninsured families, then it should have higher stress levels than those with lower percentages.
Of course, we think this methodology is riddled with faulty assumptions and fails to take many other contextual factors into account, a step that is especially crucial when dealing with subjective concepts such as stress levels, which vary from person to person.
How about your data-driven content? Do you recognize any of these common errors in your methodology? We would love to hear your comments, questions or feedback. Just drop us a line in the comments section below, and we'll get back to you.