The Future of Data-Driven Marketing

A few weeks ago at Adobe Symposium in San Francisco, my colleague and friend, marketing analytics expert Jeff Allen, gave a great talk on the future of data-driven marketing (DDM). In addition to sharing some exciting news regarding the fall release of Adobe Analytics, Jeff shared his thoughts on the evolving role of analytics in marketing, including the predictive power of the insights they produce. But what struck me most during the talk was that what he was really saying is DDM is a continuum. If you think about it, like so many things that are developmental, there was never a time when marketers weren’t data-driven to some extent (e.g., corner store owner discontinues certain food items because of poor sales performance, or perhaps stocks emergency supplies based on historical storm patterns). At the same time, DDM will never stop evolving; it is doubtful marketers will ever tap out their data, or exhaust their means to take advantage of it. Even the most ambitious companies will never feel like they’ve arrived, when it comes to the future of DDM.

I was fortunate enough to participate in the couch discussion that followed the session, and had a chance to share my own perspective on where personalization and optimization fit into this continuum. For me, DDM of course starts with observation and measurement, but quickly becomes more about insights—those gems that tell you what’s really going on! But insights are only as good as the action you can take on them—and in a timely fashion (before they cease to be meaningful). This is where data starts to really drive marketing because it has (observable and measurable) impact. It’s that virtuous loop you should be picturing in your mind right now! Most marketers believe in DDM, but readily acknowledge that the challenge is execution. For people like me and Jeff, this is where it starts to get exciting; this is where we can start talking about how we advance DDM through predictive analytics, automated decisioning/personalization, and other cool things we’re working on at Adobe.

Before we jump too far ahead I want to point out a few trends I pulled from Jeff’s talk that should resonate with digital marketers in every corner of the industry, plus what they mean for the future of optimization. Use this as a benchmark of sorts to gauge your organization today and where you’re heading tomorrow—or, even, to set your sights on something bigger and better in the year ahead.

  1. Make analytics actionable

At the root of Jeff’s presentation was that we, as marketers, need to steer analytics in a more actionable direction so we can maximize their value. Gone are the days when analytics simply served as a rear-view mirror. The past is the past, and although sometimes it’s a good indicator of the future, we have an opportunity to do better than that. The end result? Better, more relevant customer experiences. According to last year’s Adobe Digital Roadblock study, this call-to-action couldn’t come soon enough: more than three in four marketers said they need to be more data-focused to succeed.

So how to get there? Start by asking yourself a critical question: what’s the desired outcome of your marketing initiatives? Every company and every end goal is different. Define yours and work backward. My goal is to improve conversion rates, increase average order value or video views, double my email open rates—KPIs that tell me how my business is doing. Now ask yourself how data is driving the decisions you make regarding the action you take against these KPIs. Do you ignore it? Are you persuaded to do something completely contrary to what the data suggests you do? Or does the data sway, influence, inform, or maybe even drive your decisions. Perhaps you’d even go so far as to say the data makes the decision and even takes the action on behalf of the marketer. For many marketers it’s a case of how confident we are in the data, and more important, the insights we glean. Contribution analysis is a great example of gleaning insights from data; it’s the ultimate “so-what” analytics arrow in the marketer’s quiver. Knowing with confidence what specifically in your marketing mix is having impact right now could just be the multimillion dollar insight you need. It should be second nature at this point but, still, nearly half of marketers reported trusting their guts when investing marketing budgets. That means there’s still a big disconnect.

  1. Personalization should be data-driven

Personalization means different things to different marketers. I’ve met with clients who consider a geotargeted offer a powerful tactic, or product recommendations effective conversion-lifters. I’ve seen others who take it to the next level, viewing personalization through a data-driven lens. To them, every consumer interaction should be optimized, leveraging behavioral, real-time contextual, and even third-party audience data to personalize against as complete a profile as is data-drivenly possible today. Being data-driven permeates all corners of the business, informing departments and engagements far beyond the traditional bounds of marketing and analytics. Think customer service, renewals and retention, in-market journeys, and more.

No matter where you fall on the spectrum, data-driven personalization delivers—and your goal should be to make it work even harder. And consumers now expect this of you. Nearly 80 percent of shoppers expect personalization as a standard feature, with expectations ranging from remembering past purchases (59 percent) to knowing how long she’s been a customer (46 percent). And this isn’t limited to those industries we think of as digitally mature, like e-commerce and travel. Data-driven approaches are definitely part of the digital transformation the health industry is going through as well. Ninety percent of patients would share personal information if it meant better care.

  1. Data-driven marketing at its finest is automated

Marketers, it seems, aren’t born risk-takers. We have great ideas and a desire to create the next killer campaign. But all too often we get tripped up once there’s real money on the table, worrying we’ll zig when we should have zagged or concerned that life would’ve been better if we sent customers on the blue journey instead of the red. Overall 54 percent of marketers know they need to take more risks, but doubt and lack of trust in data still remain huge barriers to systematic and rigorous testing and therefore, ultimately, to optimization and personalization.

Marketers need to take smart, calculated risks that emerge from insightful experimentation, existing data, tacit knowledge, and well-articulated KPIs. They’re typically safe and controlled, and will likely be directionally what’s expected. You’ll hold back a certain amount of traffic for the control and be able to fall back on that “safe” traffic should the test ultimately not deliver.

Automated personalization through machine learning is a tactic that’s helping alleviate some of the stress marketers feel around risk. It allows marketers to relinquish some of their control over the content decisions, and as a result personalize experiences at scale to millions of individual website visitors. Last year I sat with Amy Lew, a senior UX manager at Adobe, and she talked the ins and outs of automated personalization and how it’s setting a new standard for testing and optimization. She explained that Adobe Target’s automated personalization capabilities, “can take [testing] further … by basically pairing audiences with experiences or offers based on algorithms and our own proprietary logic.”

And the risk? Although it does require marketers to put their trust in the data and the power of the algorithms, they’ll see wins right off the bat: “our personalization engine,” explained Amy, “and … data-driven, self-learning, and algorithmic approaches to targeting content can offer quick wins with less effort.” It’s safe, it combines the best of marketer knowledge and machine learning to create an added barrier of protection, and, more importantly, it reacts in real-time—no waiting for the test to see how things shook out.

Marketers simply don’t have time to watch the kettle boil these days; nor will they be able to scale their efforts if they have to continually monitor their tests and make decisions on which actions to take, and when. To address this we’re rolling out an automated testing feature this month in Adobe Target that allows marketers to walk away from a test, and rely on data to drive action in their absence. The “auto-allocate” feature will start funneling valuable live traffic to the an experience as soon as it starts to look statistically like the winner, just as a human marketer would if he or she were able to make the same observation!

Where do we go from here?

So just why aren’t all marketers jumping on board the DDM train? According to the Adobe Digital Roadblock Study, nearly two thirds of marketers said they’re comfortable adopting new technology once it’s mainstream. But waiting for others to adopt automated personalization and other DDM tactics isn’t the way to go. We’re past the early adopter phase and into a meaningful proliferation moment in time. Wait and you’ll miss out on countless benefits, relationships and incremental revenue sources.

Data-driven marketing isn’t going anywhere—in fact, it’s only going to get bigger, better and more integrated into consumers’ experiences over the coming years. Consumers crave spot-on experiences and we, as marketers, want to deliver. Marketing analytics, data science, and machine learning are all components in this push towards a more “always-on” environment where experiences and expectations go hand-in-hand no matter the brand, the customer or industry. It’s going to take top down buy in and, above all, more actionable data that informs and inspires every movement we make. The future of data-driven marketing is now.