Personalization’s deep data foundations

Data-driven personalization is the Holy Grail of marketing today. Various studies, mostly sponsored by retail-related organizations, have claimed that between 65-and-95 percent of customers prefer personalized experiences. These aren’t just “entitled” customers who expect everything dished up to their individual liking with increasing levels of speed and convenience. It’s every customer who has to devote hours wading through a sea of irrelevant information and ads to find what they’re looking for.

What a wonderful, serendipitous relief for someone to just drop the desired thing, produced by a company the customer feels good about, right in front of their faces! That’s the goal of personalization.

But today the practice often falls far short of the goal, which is why consumers get ads for things they’ve already bought, or decided against, or that are only peripherally related to something they had a fleeting interest in.

To actually create a magic moment of right-thing-right-time is both a technological feat of data collection and analysis, and the product of an insightful marketing team who knows how to tell its artificial intelligence (AI) what information to fish from that ever-expanding sea of data.

[Related read: Customer entitlement: the high price we all pay]

The personalization-privacy paradox

If delivering spot-on personalized experiences is the magic moment, however, the process of data collection that’s necessary to empower it might most appropriately be called “making the sausage.”

What a wonderful, serendipitous relief for someone to just drop the desired thing, produced by a company the customer feels good about, right in front of their faces!

Nobody likes thinking about others tracking their behaviors, it just feels creepy. But they do like personalized offers and service. That’s the personalization-privacy paradox. Research shows that customers unconsciously perform a mathematical calculation to decide which one is more important to them. If the data being collected is low-risk, and the personalization experience that results is perceived as high-value, the relative cost of privacy seems small.

For example, if a customer is shopping for a pair of riding boots for winter and they are served up the exact pair they want at a great price, they have given up little privacy for a great benefit.

If, however, the customer searches for information about an embarrassing health issue, research shows that the risk of having their privacy violated does not cancel out because of the value of the offer.

One example of privacy violation is when Target accidentally broke the news of a teen girl’s pregnancy to her father by sending “personalized” offers to her home.

When people think about having a personalized experience, they don’t want to think it came from crawlers digging around in their information as they search the internet. But it does. So companies had better make their personalization efforts worth that loss of privacy.

To actually create a magic moment of right-thing-right-time is both a technological feat of data collection and analysis, and the product of an insightful marketing team who knows how to tell its AI what information to fish from that ever-expanding sea of data.

Most customers understand that “cookies” are “tracking” them. They know their data is being collected spread far and wide, and this contributes to the impression companies are peering into their homes. In the case of Alexa that may be true, but generally it’s not how brands create personalized experiences.

[Related read: Did GDPR kill the personalization movement?]

Breaking it down: Personalization in layman’s terms

This is how it actually works: Customers who visit a specific site agree that the site can store “cookies” on their computer. These are bits of code that track everything the customer does on that site. Cookies may store the customer’s passwords, credit card information, transaction history, search history, and this makes it easier for a company to respond to a customer’s preferences when they visit that site. Data collected by the company about their customer is called first-party data and companies may use it to provide more context within their customer relationship management software.

Yet when customers click “agree” on that cookie policy, they may not realize that they just gave permission for the company to share their information with the company’s partners, too. This is outlined in the agreement, though few people read it. The first company won’t share private information—like passwords. But they will share information that helps other companies sell to the customer, things like search and transaction history. Partner companies use this data to target the customer with ads. This is called second-party data and may be stored on a data platform that combines first- and second-party data , to create a more complete picture of the customer. Personalization, in other words, can only be effective when there is enough data to leverage.

Data shared by companies doesn’t identify the customer as an individual. Instead, customers fall into a type or category that partner companies might be able to market to. For example, if a customer books an ecotourism trip with Green Tours, Green Tours’ partners might recognize the customer as an “ecotourism enthusiast” or “green travel enthusiast.” Marketing automation software is then set up to offer specific ads to customers who fit this category. The customer might find they’re suddenly getting ads for a new backpack from a company called Outside Gear. They might also notice offers for a raincoat, travel insurance, a stay at an ecotourist hotel from other companies. AI used by each company will collect and analyze data from previous transactions with other customers, as well as data shared by Green Tours, to predict which offers are most relevant.

Picky about partnering

Now, Green Tours is likely to be particular about its partners. They are less likely to partner with a cheese company, for example, or a fashion brand that sells expensive purses. So while offers to the customer are unsolicited, they are also somewhat personalized to appeal to the customer. If the customer does buy a backpack from Outside Gear, that company collects data as well, and shares it with their partners.

Every purchase provides vast amounts of data about purchasing habits. Companies can also augment their data with third-party data from other companies willing to sell it. These data providers use many sources, including social media, to find out about the customer’s age, marital status, job, political affiliations, type of home, type of transportation, tendency to volunteer, food preferences, medical conditions (from over-the-counter purchases), and so forth.

Personalization, in other words, can only be effective when there is enough data to leverage.

Outside Gear might buy third-party data on its customers, but that data is delivered as an aggregate of information that helps with segmentation: For example, “30% of your upper-end backpack buyers are between 35-50 years of age.” They will not get data that can identify any one person.

The Forbes Insights and Arm Treasure Data survey discovered that the channels companies use most to collect data are mobile apps and email, though email is considered a low-value channel for enterprises. Customer loyalty programs and mobile apps are considered the highest value. Now imagine this data collection is done every day, on every transaction for years, and shared. It’s pretty easy to get a picture of who you’re dealing with.

The role analysts play

The thing is, some of this information is helpful, and some isn’t. It takes analysts to figure out which data is worth considering when creating personalized experiences.

Historically, customers were segmented by age, gender, race, location, income, and other factors. But with better technology, a MarketingWeek study revealed that behavioral data is far more helpful for creating the kind of personalized experiences that increase revenue. Respondents said behavior, location, personal interests, life stage, and attitude are some of the most important forms of segmentation. French food company Danone segmented its customers into “tribes” based on certain passion points and saw a 40 percent lift in ad recall, the article said.

[Related read: Thanks to emotional connectivity, I’m a proud member of the Spotify tribe]

Jean-Marc Bellaiche, chief strategy officer at experience analytics platform Contentsquare said companies need to be as unique as this to understand their customer segments. “The companies that are thriving in this new era of experience are the ones which have understood that there are no real best practices, there is no cookie-cutter solution to the experience challenge,” he said.

Respondents said behavior, location, personal interests, life stage, and attitude are some of the most important forms of segmentation.

A Gartner report, The Essential Guide to Marketing Personalization, noted that Clorox had been collecting various data on a specific marketing segment for its pet mess products. They include the customer’s age, cat ownership, address, type of home, type of transportation, frequency of grocery trips, and several other things. The company wanted to stop wasting money on mass personalization efforts. So their analysts began finding ways to hone the data. They decided that one of the most important factors was the accuracy of a metric and only used data from their most accurate sources. This cut out certain data points like whether a customer lived in a walk-up apartment or what their preferred mode of transportation was. The team then went through other criteria, such as Reach, meaning which dimension should they include because it would yield the largest audience. Each criterion reduced the number of data points that they included in their personalization efforts. The result was a big savings in marketing spend and a five percent increase in sales in one year.

Better experiences, more revenue

“Our customers have many different identities, and they are creating ever-growing amounts of data everywhere—both inside and outside our company walls,” said Zendesk CEO Mikkel Svane in a blog post. “To make matters worse, legacy CRM platforms force us into their proprietary technology. That makes it difficult, expensive, and nearly impossible to see the many different dimensions of customers and their data—a view that’s required in order to make meaningful improvements to the customer experience.”

[Related read: To excel at customer intimacy, you will need data]

With personalization on center stage in many ways, more and more technologies are being developed to make it easier to know what customers want. For example, there’s a move toward a new kind of open, flexible CRM that allows companies to connect customer data, wherever it lives, so that it can be used across the business.

Likely, these technologies will also increasingly respect the privacy that customers’ require. Most companies are just beginning to explore how to do this well, but the foundation for any personalization effort has to be having access to the technology that allows you to collect and use data effectively.

The final result is the ability to give unique segments of customers magic moments. But magic moments are made of data.