Billion-Dollar Baby: Customer Data Onboarding

Last week, Bruce Biegel, senior managing director of Winterberry Group, led a Direct Marketing Club of New York discussion on “customer data onboarding.” It’s the process of linking offline data with online attributes (cookies, IP address, device IDs, non-cookie identifiers, among other identifiers) in order to perform any number of marketing use cases, what often is referred to as “data activation.”

This blog post opens with a 2014 local TEDx talk of Dr. Charles Stryker, who left us unexpectedly last month. To say the least, “Charlie’s” legacy as a data innovator, business leader and financier is very much alive, well and profound.

Dr. Stryker truly set a stage for what has transpired three years hence — and likely decades to come.

Last week, Bruce Biegel, senior managing director of Winterberry Group, led a Direct Marketing Club of New York discussion on “customer data onboarding.” It’s the process of linking offline data with online attributes (cookies, IP address, device IDs, non-cookie identifiers, among other identifiers) in order to perform any number of marketing use cases, what often is referred to as “data activation.”

These use cases for customer data onboarding most often include (but are not limited to) digital display targeting, “walled garden” targeting (such as ad targeting inside social media platforms), consumer analytics and insights, measurement and attribution, site personalization, and addressable television and online video targeting. (Readers, please don’t confuse customer data onboarding with customer onboarding — they are not interchangeable.)

Winterberry Group estimates this market to exceed $300 million today — but may eclipse $1 billion by 2020.

“Consumers are constantly connected,” Biegel said, noting each connection can generates volume and variety of data. “Today, the average consumer uses close to eight connected devices per day. One of my colleagues counts 22 connected devices!” Such connectivity generated close to 44 zettabytes of data in 2016 — and should more than quadruple that data generation to 180 zettabytes by 2020.

How do we structure these data to help recognize that customer from device to device in privacy responsible ways — to enable all these vital use cases? The answer, customer data onboarding.

Certainly, the big driver of customer data onboarding is brand engagement — all those use cases demanding customer identification and data quality to enable 1:1 prospect and customer communication. First-party data (a brand’s own data on its customers) needs supplemental third-party data (read, ad tech and data providers) to complete the customer view across connected devices, enhanced by offline data (often transactions).

Joining Biegel were panelists Anneka Gupta, co-president, LiveRamp; Kevin Whitcher, VP for product, Oracle; and Paul Chachko, CEO, Throtle — all of them representatives of distinct industry players in the customer data onboarding space.

In onboarding’s nascent days, according to the panelists, brands may have matched 5 percent to 10 percent of their offline-digital customer file — but today, such match rates have grown to 50 percent or more, depending on the device (mobile versus laptop, for instance), helped by more deterministic (certain match) and probabilistic (likely match) means for identifying unique individuals and households.

“Never trust a vendor who claims a mobile match close to 95 percent,” Chachko said. “Probably the best upper match rate we have [in mobile] is 75 percent.”

With individuals changing devices all the time, adding new ones and retiring old ones, moving inside and outside login environments, adding new email addresses while not necessarily discarding old ones — all of these data points help to inform a “consumer ID graph” that is increasingly person-specific (rather than household), according to the panel.

“It used to be [data onboarding companies] used these graphs to facilitate the data matching inside their operations,” Oracle’s Whitcher said. “Now more and more clients want to access and use the graphs themselves — it’s a new product offering in and of itself. … It all comes down to what does the brand want to do? Digital display targeting is the most common use case — but more marketing uses are emerging, such as attribution.” Mixed media attribution models are becoming more common.

In omnichannel marketing environments, CMO’s know that last-touch attribution falls short when consumer paths to purchase are so varied.

LiveRamp’s Gupta noted that addressable television is also gaining advertiser attention, but online video targeting is a very hot use case now.

As bullish panelists are toward customer data onboarding’s growth, not all brands are ready for the discipline of such data enhancement. “Global brands must worry about privacy rules, particularly in Europe,” Whitcher said, where European Union General Data Protection Regulation and ePrivacy Regulation loom.

Closer to home, many enterprises are not yet realigned around customer experience. Data siloes by channel, and by product, interfere with a full customer view — and these siloes must be broken before data integration and quality regimes — and the insights they produce — can be applied.

“Retail and financial services are perhaps the most mature adopters of customer data onboarding, largely because they were the earliest to invest in whole-customer views,” Gupta said.

Whitcher concurred on retail — noting that competing with Amazon has spurred all types of data innovations. He also noted that auto, travel, media and entertainment, and consumer packaged goods are also fast-rising in data onboarding use. “Not all sophisticated marketers are sophisticated data users,” he said.

“When we see clients fail, it’s because they bite off more than they can chew,” Gupta said. Data science requires testing, proof of concepts, data quality and rollout testing — all practices very familiar to traditional direct marketers, but not necessarily mass-market brands.

Still, the billion-dollar market is gaining steam — just like Charlie said it would.

Note: Also see from January 2017, “Customer Data Onboarding: Winterberry Group Publishes State of Market.”]

People-Based Marketing: Targeting People, Not Cookies

With the promise of big data being able to provide true people-based marketing, the yearly spend for companies continues to increase. But is your data really as good as you think? Are you truly delivering the right message at the right time to the right person?

the pulse blog

Check out even more about personalization and artificial intelligence with FUSE Enterprise.

In 2016, eMarketer reports approximately $16.2 billion was spent on digital media. With the promise of big data being able to provide true people-based marketing, the yearly spend for companies continues to increase. But is your data really as good as you think? Are you truly delivering the right message at the right time to the right person? The overpromise of big data being able to provide marketers with this solution can create waste — both with money and time. Delivering on people-based marketing, though efficient and lucrative when done well, is turning out to be a lot harder than it sounds.

Cookies Are an Issue

Cookies have a lot of shortcomings when trying to map back to a specific individual. Cookies aren’t persistent. They can be deleted or blocked, and in most cases an individual can have multiple cookies and even have them assigned to the same device. Several advertisers have started to invest in solutions that leverage machine learning and artificial intelligence (AI) to map cookies back to an individual. But in their quest, they have forgotten the whole point of a cookie, which is supposed to be anonymous and stay anonymous. Yes, Google knows which double-click cookie maps to your Gmail account, but they’ll never share that information. That would defeat the whole purpose of why anonymous third-party cookies were created in the first place.image_1

Finding the Solution

Facebook takes credit for coining the term, and they are definitely leading the pack toward people-based marketing. The key to Facebook’s success starts with the individual versus with the cookie — they map a person’s devices and cookies back to the individual. Not the other way around. It’s a simple but important distinction, and it’s key to their success. This particular methodology works great for Facebook. After all, they already have an exhaustive list of individuals. But what about the average marketer? While Facebook will let you leverage some of what they know on their platform for a price, they definitely aren’t sharing that data. The good news is most marketers already have extensive CRM databases of their existing customers. The key is to unlock that data and target those individuals and other look-alike prospects. While it’s simple in principle, it’s challenging to put into practice.

 

image_2Learn even more about the convergence of technology and branded content at the FUSE Enterprise summit. Artificial intelligence and personalization will be featured among many other techniques and technologies.