Home » Preparing for the Future of Data: Implications of Privacy, AI, and Machine Learning

Preparing for the Future of Data: Implications of Privacy, AI, and Machine Learning

Daniel Eisenhut
Daniel Eisenhut , VP, Services & Support, Emarsys

We’re in a pivotal place with regard to customer data, and a new chapter is being unveiled — everything sort of hangs in the balance, right before our eyes. 

With so much happening recently in the way of data privacy, marketers and media companies  both face a lot of tough questions.

Facebook — and partner brands — are forced to react to the data giant’s decision to remove third party data sources from its Custom Audiences program amidst news of its misuse of customer data brought forth by Cambridge Analytica.

There’s no doubt that data and how we can use it is going to change — between this news with Facebook and new regulations like the GDPR going into effect shortly — and we’ll have to adjust.

In a lot of ways, the future of data is still quite up in the air. But what do we need to be thinking about, especially with respect to first-party data that you collect and own, in order to get ready for the future?

Who Owns All This Data? For Now, 5 Companies Hold Most of it… But That’s Changing

▬▬▬▬

For now, the majority of data is owned and operated by five companies — Apple, Facebook, Google, Microsoft, and Amazon.

But this will change in the coming years.

We’re moving toward a time and place where any and every brand will have the technical ability to collect data from all possible online and offline sources where they have touch points directly with customers.

The place where I see most companies that struggle is in unifying various data points from separate sources into one consumer data platform — all while attempting to offer all marketing business units (email, social, mobile, web, print) one “source of truth” when it comes to identifying and determining who to target.

This is a major challenge that prohibits a cohesive experience across channels for users. It also blinds the marketer — if you’re only able to look at data points for individual channels, you’ll never see the whole picture.

“Most #marketing teams struggle to unify all channels from a “single source of truth” which impedes the #CX & also blinds the marketer,” says @eisenhut_dan                            CLICK TO TWEET

How Much Data Do You Really Need to Make Intelligent Decisions and Gain Insights? It Depends.

▬▬▬▬

“…You may not need all that much data to start making productive use of machine learning. The performance of most machine learning systems improves as they’re given more data to work with, so it seems logical to conclude that the company with the most data will win… But if success is defined as significantly improving performance, then sufficient data is often surprisingly easy to obtain.” Harvard Business Review, 2017

The first step in thinking about how much data you really need is understanding your ideal customer journey.

Where are your customers engaging with your brand and product (channel), and how is he/she engaging with you (interaction in the channel)? These two simple questions give you the best starting point to analyze and answer. You should also ask:

  • Which channels are you prioritizing to drive traffic from?
  • Which channels are driving consecutive product and content views?
  • Which channels drives profitable and repeatable conversions (in order guide proper retention marketing)?

Once the consumer journey and all touchpoints are defined, determine which data points you want to use from those interactions. This could be a sign up with a preference center to better understand leads, a mobile phone number at checkout in order to support a mobile shopping club, a product view to drive relevant content, dates when these events occur, and more.

Once the sources and interactions per source are identified, this is, in my opinion, a solid foundation for machine learning products to start making predictions about who the right audience is, what content to serve to this audience, when to serve the audience with relevant content, and which channel is ideal for content.

We’re on the Cusp of Mass Adoption of AI and ML — What’s to Come?

▬▬▬▬

When it comes to practical application of AI in a marketing sense, we’re starting to see more and more marketing teams using AI for personalization, promotion, procedural efficiency, and more.

Ideally, marketing organizations will use AI to spend the majority of their time creating business strategies and relevant content to support these business strategies. The machine will take it from there to reach desired audiences with timely placement of content per channel and per individual.

The reality today is that most AI systems are just basic logic-based systems guided by human programming. Beyond basic pattern recognition, most AI systems aren’t truly “smart.” But with the democratization of data we discussed above, could this change ?

Machine learning is dependant on human input of data, as at this stage humans still define the business strategy and how to get there.

How might we use AI to drive business? You can drive more revenue from, say, first time buyers or win back more defecting buyers by using AI to optimize content and offer for each person.

The marketer will define the audience, the time of trigger, the content, and the channels, but the machine will execute. In the near future, AI will start learning each area of manual steering and the marketer can start to replace one element at any given time (e.g., AI should define the right channel for Content A to be displayed to this Customer B, or which content to display to a given audience in a certain channel).

As we start to replace human decisions for who to target, when to send, what to offer, and on which channel, AI will start to be strategy-driven and marketers can mainly focus on defining new strategies — without having to focus on the tactical execution of that strategy. That will be the real promise of AI — and the data which drives it — and is what I’m most excited to see in the next several years. ◾

I’ll be publishing a second “Future of Data” piece soon where I’ll share thoughts on new data-related jobs that will emerge in the marketing organizations of tomorrow, big data, and its implications for both blockchain and VR/AR.

Handpicked Related Content: 

 

Marketing is in a constant state of change. Are you keeping up? No sweat — we’ve got you covered. We’re revealing 5 big data predictions for the next 5 years, straight from the desk of the CMO: “5X5: 5 Marketing Predictions for the Next 5 Years.”

New Call-to-action


Daniel Eisenhut is Vice President of Services and Support at Emarsys, and has been with the company since 2011.

Connect with Daniel: LinkedInEmail