COMPREHENSIVE GUIDE TO CONSUMER DATA SOURCES
By Ami Murphy Iannone, Creative Director - January 24, 2019
The types of consumer data that brands are looking for and how to evaluate them
We’ve established that both brands and agencies are moving toward data-driven marketing in this new digital world. There is no turning back to the days of vague metrics or approximated audiences being “good enough.” The competition is too steep and the margin for error is slim when everyone has access to the same data from the social channels— they’re all trying to differentiate themselves within the same scant data points.
So brands are increasingly looking for richer sources of data to drive their decisions and help them separate themselves from the pack.
Brands want rich, predictive data about their consumers
To cut through the noise of today’s endless stream of content, marketers are turning to data science and artificial intelligence to learn more about their consumers and create content tailored to their interests. This has perpetuated a data arms race based on which insights can be most impactful for marketers. The challenge for marketers is that not all data is created equal.
So, as a marketer, what data is most relevant for your strategy?
In evaluating which audience insights are relevant, marketers need data that is both holistically informative and predictive. The data needs to be informative in that it is an accurate representation of the audience you’re analyzing. When this is done holistically, the data will encompass a wide range of metrics to analyze. Predictive data will not only be representative but will also provide the type of value that can lead to your “a-ha moment”.
Qualities of Data
The most desirable types of data can give a complete picture of the consumer and their lifestyle to provide marketers with the best understanding of that consumer’s needs. The most desirable qualities of data are context, demographics, behavior, and sentiment — ideally, each of these qualities would be acquired real-time so that the data never gets stale.
Contextual data tells you about the conditions of the world around a consumer at the time that they take an action.
Demographic data tells you about the consumer themselves, including age, gender, location, income bracket, and education level.
Sentiment data tells you how the consumer feels about certain topics or products.
Behavioral data can give you insight into their preferences, interests, and habits.
Real-time data tells you about the current realities of your consumers. It is quickly accessible and constantly refreshing.
Predictive data provides you with the ability to forecast an outcome based on the historical actions of the data set.
Sources of Data
Not all data sources are created equal. Let’s take a look at a few of the different types of data available to marketers and the advantages and limitations of each of these sources.
Polling and Surveys
Traditional polling techniques include mail, internet, phone, and in-person interviews as well as survey collection through the same avenues. Multiple modes of response collection may be required to reach the desired demographic.
First person response to your brand’s questions from a targeted group of respondents. This can allow you to gather direct feedback from a certain demographic that you’re hoping to learn more about. Additionally, surveys are relatively inexpensive because there are many companies and technology platforms that power this kind of information gathering.
Surveys require rigorous methodology and execution in order to ensure that the data returned is unbiased and viable for your brand’s applications.
Surveys require both participation and honesty. These two factors can make it especially difficult to get a whole, representative picture of your audience from survey data. Survey participation has been dramatically waning over the years which makes it more and more difficult to get an accurate and representative sample set. Pew Research Center outlines the challenges the survey industry has been facing:
As Americans are now faced with more demands on their time, they are exercising more choice over when and how they can be contacted.[...] As a consequence, response rates have continued to decline over the past decade.
These changes mean that it now takes much more work and time to conduct a viable survey. Furthermore, the information gathered through polls and surveys is self-professed which means that it can be colored by a person’s desire to represent themselves a certain way.
Web History and Cookies
Ability to track a website visitor across the web to view the other sites they visit and deliver them relevant ads across their browsing experience.
Allows you to build a picture of the consumer based on their other interests beyond your site. Cookies provide a great source of behavioral data. Cookies allow you to track a user across the web and deliver retargeted ads to them, increasing the number of times they’re exposed to your messaging.
This type of data only represents behavior and does not encapsulate sentiment or demographic information about site visitors. This data source must be coupled with other data sources in order to provide a complete picture of the target consumer.
Users are able to clear their cookies removing them from your radar — and many people do as the current data climate has made consumers reticent about cookies and other overt forms of data tracking.
Smart Devices & Apps
Data collected from devices with deep infiltration of consumers’ everyday lives, like voice-activated smart home devices and smartphone apps.
Apps and smart devices provide a rich source of behavioral data and are unique to these other datasets because of their intimate place within a consumer’s life — these devices are integral to consumer’s home lives and are carried around with users everywhere they go. You can track all of your customers’ moves within your app to deeply understand what content and/or products are most interesting to them and where they spend the most time interacting. You can use this information to provide super-targeted messages.
To obtain this rich data, you must either have an app or purchase it from third-party sources, which can be pricey. To gather this info yourself, consumers must download and engage with your app. Getting users to download an app is an additional marketing hurdle alongside the challenge of maintaining a seamless and useful app experience for customers.
In fact, this data may be a little too precise. Today’s data climate has made many users aware of and hesitant about the data exposure that apps entail. Users are becoming more aware of the intrusive nature of this kind of tracking and resentful of the ‘creep factor’ associated with it. A recent piece by The New York Times revealed that a single woman’s location was tracked and recorded about every 21 minutes. The Times was able to easily identify her and follow her to work, the gym, a doctor’s appointment, an ex-boyfriend’s house and even track how long she remained in each place — a revelation rightly discomforting as each person reading this likely has a smartphone in their pocket.
Correlating earned email lists with third-party data to deepen the understanding of your audience.
Many consumers use their email address as their ticket to unlock the entire internet which means that you can obtain much deeper information about an individual consumer once you have that identifier.
Building a robust email list isn’t easy. To collect email addresses in an ethical way, you must earn the trust of users and provide a compelling value exchange.
Once you’ve collected those emails you then have to execute the perfect balance of email correspondence to maintain the list without abusing their inboxes and making them unsubscribe. People are often wary of giving away their email addresses— aware that they are a form of digital currency these days— and may have a ‘junk’ address that they use to sign up for offers or promotions.
If you’ve been successful in establishing, maintaining, and growing an email list, you’re still not home free. Email addresses themselves are not especially useful as audience intelligence tools. These lists require third-party enrichment to become signifiers of user behavior across the larger web.
Information from social media platforms, like Facebook and Twitter, about an audience’s makeup and interaction with a brand’s social presence.
Social media data can give you a first-person account from your consumers. Social media data allows you to analyze the way that your engaged audience interacts with your content as well as their interests and brand affinities based on their social behavior. Because social media is ‘always on,’ this data source is constantly refreshing.
First, you’ve got to have followers to get the analytics. This requires a certain amount of work to build and maintain an engaged social presence across different networks. For most brands, this is a challenge. For agencies, this requires access to the brand’s owned social channels which many are reticent to provide to external partners due to brand safety and privacy concerns.
Social metrics only allow you to analyze the actions of the people who choose to engage with your brand — a limited, captive audience and a potentially saturated one. Those engagement metrics are historically thin. It can be difficult to properly weigh the value of a like, share, retweet, reply, or comment as it pertains to ROI.
Social data can be deceptive. Even if users interact with content in a negative way, as trolls do, the social platform still registers that interaction as an indicator of interest without taking sentiment into account. You’ve got to pay close attention to the micro-level interactions to make sure that a post is performing well and not going viral for negative reasons. Beyond the human behavior issues, there are the challenges of bots and fraudulent engagement metrics to skew this data.
Joseph Harper, the social media lead for Kellogg’s UK, discussed the unique challenges of social media engagement through the lens of influencer marketing at the Digiday Brand Summit in November 2018. Harper noted the ways that social data can be easily skewed and manipulated, including engagement stats:
“One agency we work with said a campaign was a success because it generated loads of comments, but when we dug deeper into the report, we realized that the influencers we’d paid had just gone to a WhatsApp group of other influencers and asked them to make all of those comments.”
For social media data to be a truly meaningful glimpse into your audience’s lives, it should be coupled with technology solutions or serious man-hours digging into the sentiment and context around the engagement.
An estimated 80% of all internet traffic will be video by 2019. With all this video being watched across social and web, viewers are leaving metadata trails that are rich in informative insights for marketers to analyze.
There’s lots of this data available and it encompasses several important qualities, including contextual, sentiment, and demographic.
Social video data is an amazing source of real-time data of what people are watching and what people are saying about a specific topic. This can just be traditional social listening, but it can be a lot deeper looking at specific engagement with videos. Machine learning and artificial intelligence can help determine which data points are actually useful and predictive and which data points to ignore.
According to Nielsen, U.S. consumers watch nearly three hours of video on their computers or smart devices each week. The time spent viewing social videos is what makes this data so valuable and predictive. Time is the ultimate resource. We all have the same amount, so where you spend your time shows a lot about your interests and priorities. Basically, you are what you watch. Viewers spending their time on video are giving strong, predictive indications of what they like. Listening to social video data lets you drag in a wide net of information about an audience and A.I. can help decipher what pieces of that huge amount of data are actually relevant.
Requires a robust amount of video content on a topic or channel in order to analyze. However, this limitation is not unique to video. Any valuable dataset must be large enough to have statistical significance. The unique aspect of this challenge would be around clients who have not yet developed a substantial owned video presence. You can still analyze the videos of competitors and specific topics in order to understand and implement the value of video data for a client without a large video presence of their own.
Brands want better audience understanding with context
There are tons of sources of data available to marketers today. All of this information is useful as a tool to paint a larger picture of your target consumers’ lives and allow you to find ways to provide them with more value. But data is only valuable within context and data can be deceiving.
If you looked only at the performance metrics for Pepsi’s ill-fated Kendall Jenner campaign without context, you might think it was a raging success. Without understanding the audience sentiment and the sociological context you would only see that viewers flocked to watch the ad and commented and shared like crazy. This demonstrates the importance of sentiment as a vital part of all truly useful data sets because it can help provide context and inform future actions.
Data is only valuable for what it can tell you and what actionable steps it leads you to.
The best data leads marketers to a deeper understanding of their audience.
As the future of video production and consumption changes, so too does the future of advertising personalization. Consumers don’t just expect entertainment, but they now expect on-demand, personalized experiences. Artificial intelligence is moving the industry towards a future (begun with A/B testing) where each message is succinctly and perfectly tailored to the individual viewer based on that person’s digital behaviors.
Brands want faster, smarter decisions and better results
No matter the source, there is an incredible amount of data available to marketers. The number of owned digital touch points available for marketers to analyze, not to mention the availability of third-party data for purchase, is overwhelming. Having too much data can be paralyzing and if you’re unable to sort through that data and quickly find actionable insights, then the data will get stale and become useless.
Artificial intelligence is best suited to power this new, necessary kind of audience understanding because it can keep up with the speed of creation. The aforementioned study from Magna Global sites an incredible statistic: The digital universe grows by 2.5 Exabytes every day — that is equivalent to 90 years worth of HD video produced Every. Single. Day. There is just no way that humans could collect or process the magnitude of the information available. But pointing artificial intelligence and machine learning at those massive troves of data can help humans digest that staggering amount of information and detect patterns. Allowing humans to make faster, smarter decisions rooted in deep understanding.
The ANA’s study on in-house agencies cited “speed” and “nimbleness” among the top benefits of an in-house agency versus a traditional external agency relationship. Many traditional sources of data can be laborious and time-consuming to collect, process, and turn around to clients packaged as thought leadership. Accessing the correct technologies to collect and correctly interpret data is vital to agencies’ continued competitiveness.
Agencies have higher stakes in presenting correct information quickly. As insourcing continues and budgets for external services tighten, brands will continue to scrutinize the creative and strategic decisions their agency partners make on their behalf and demand not only quick and data-backed proposals but also an accurate measurement of success.
Brands want more accurate reporting
Alongside the incredible availability of data sources comes the incredible responsibility of reporting on campaigns. Like it or not, surface-level metrics aren’t going to cut it anymore.
As Kellogg’s Harper noted, “We’re trying to move away from being solely reliant on vanity metrics and take into account the sentiment of posts as well as the different types of conversations happening around them.” Many brands are following suit.
Agencies will need to begin to employ smarter tools to help them measure and report on the nuances of campaign activity beyond the surface-level performance metrics.
Artificially intelligent platforms that can help identify sentiment and provide predictive reporting will help agencies distinguish themselves as serious data leaders.