Digital analytics today is burdened by disillusionment and disappointment. Not that there are no success stories with digital analytics. There certainly are. But they are comparatively rare. Much more common are legions of valiant but frustrated marketers continuing to struggle with the basics:
Is data collection accurate? Once we learn what the data tells us about our business, are we in a position to do something about it? What happens when our agency tells us they’ve taken care of measurement and, behold, the campaigns are “all good” (or at least not a total waste)? What does change really look like, and can we make it happen in time to matter? How do we do that without automation? And where are the successful predictive models that drive automated responses?
The unanswered questions don’t stop there, but for the sake of brevity we shall.
No one suggests that organizations go without analytics. And many businesses do get to a place where they are comfortable measuring with accuracy and understanding. Many fewer end up being able to fix any but the most egregious “disconnects” between themselves and their customers. The vast majority settle for knowing what happened, with a moderately strong determination to do something about it “in the next release.”
How Digital Can Deliver for Marketers
Many of the most dire threats to success in digital marketing can be overcome by adhering to a process. The process is not very mysterious, and, in fact, can, with some alteration, be applied to almost any endeavor requiring rigor and results.
The process looks something like this:
1. Determine key performance indicators
2. Implement data collection and reporting
3. Review and analyze reports
4. Make content changes
5. Measure again to prove success
Following these will go a long way to avoiding disappointment and marketing paralysis, but often it proves devilish hard to get through the process.
Saved by Automation?
The toughest parts of the above process are numbers three and four.
It’s easy enough to figure out your basic metrics and get the data collected properly as long as you have a team of analytics experts. We’re pretty much overrun these days with analysts, but often it’s tough to turn what they say into recommendations. Then, the most difficult part is getting changes made. Figuring out what changes to make, and how to get them made, typically slows the process nearly to a halt.
Automation will be key in changing this from a roadblock to a starting block.
With Tealium’s AudienceStream, you can build in rules and thresholds that send out directives to content delivery systems that let you know it’s time to contact the customer with an offer (for instance). The key to its success is its timeliness and the certainty of its execution. It becomes automatic.
Conductrics deploys Artificial Intelligence to create a system of learning and action based on data. For marketers, this means that Conductrics will facilitate the creation of an “agent” that seeks out challenges and then tackles them (for example, it looks for meaningful patterns and then can direct content to be distributed as needed). Conductrics has likened their agent to a Roomba for digital analytics. It learns its environment and then focuses on doing one task very, very well automatically.
XplusOne [x + 1] markets a product called Origin. According to the firm, “Origin harnesses data to drive real-time, one-to-one interactions across all your digital channels, so every prospect and customer interaction is more relevant.” They also deploy a Data Management Platform that controls numerous customer touchpoints automatically.
These products help conquer the challenges of what many today call omni-channel marketing. They help address how customers can be reached in various “states,” as Rand Schulman has pointed out.
Automation is moving ahead rapidly. It may save analytics by embedding it into an automated process — which probably is where it belongs.
Last week I had the opportunity to lecture in the content engineering class at the University of the Pacific (UOP), where I’m an executive-in-residence for digital marketing and media and where I have helped design the curriculum. For the last few years, I lecture every semester on the newest technology-drivers
What a difference a few years make and I am pleased to say that change is afoot at the university and across the U.S., as schools are now building programs geared to meet the critical requirements of business in some of the most sought-after areas of marketing: social media, content marketing, and content analytics, where clearly an imbalance exists between skills taught in classrooms and the skills sought in the marketplace. A few years ago many of the students questioned the value of analytics.
This time I presented my case study of a consumer packaged goods (CPG) company using cross-channel analytics, correlating online and retail behavior, utilizing mobile apps, iBeacons, and augmented reality. I’ve written about it here. In class we mapped out just where content was created, who created it, and where measurements are needed. We talked about CPG company goal of increasing lifetime value (LTV), greater customer engagement, and we designed calls to action and discussed and identified conversion events and stages. We talked about metrics for augmented reality, profile information, and purchase history and their relationship and how we look to normalize structured and unstructured information in analytic applications. We discussed what person in the process is responsible for what action. They get it.
Universities and their business partners are indeed looking for ways to leverage their current curriculum toward a higher technological quotient. They see the need to produce graduates with the skills geared for today’s competitive environment. Yet teaching “Internet marketing” and “how to” classes – most often a minor repackaging of traditional marketing – is only a quick fix when instead we need an earth-moving overhaul at a foundational level.
For the last four years, UOP has been creating a program that addresses the needs of graduates, one that leverages the university’s existing curriculum in business, liberal arts, and engineering, but also remains adaptable to train students in new competencies as the market changes dynamically. When we started the program, few students in the liberal arts school studied the scientific method of test and control or used analytic tools. Today, students in the content engineering course build websites and use analytic tools to test and optimize the results. Some sites have thousands of visitors a day. They are learning that creativity without conversion equals zero.
We need to create a stronger culture of measurement in higher education, one that is market-based and rewards innovation. Anthony P. Carnevale, director of The Center on Education and the Workforce at Georgetown University, has called for a fundamental shift in thinking about the way students are educated. He writes, “The old model, where you go to college and then go out and find a job, is largely outmoded. It needs to be replaced with a new model, in which college years are spent explicitly preparing for an occupation.”
U.S. higher education – long a source of pride and differentiation across the globe – is undergoing a true crisis of value and identity. Pundits wonder whether universities are the next “bubble” of the U.S. economy, and university students are questioning whether their high-priced education and gargantuan debt loads – up more than 50 percent after inflation from a dozen years ago – will position them for a college-worthy career.
Says Jim Sterne, chairman of the Digital Analytics Association, “The need for analysts and content engineers who can determine the value of content is so great that our association’s online courses have been steadily sold out since inception in 2006. When times are good, companies invest in tools and systems.”
The market needs well-trained content marketers who can create compelling content and measure and optimize that content using new analytics, predictive modeling, business intelligence (BI), marketing, and content management tools.
Marketing students should be given rigorous, cross-disciplinary training in writing, analytics, and technology; engineering students should be taught to create content; and English, journalism, and communications students should be taught about optimizing content for business value.
“I’m pioneering content engineering in our English department focusing on teaching the latest tools of analysis, analytics, and optimization alongside traditional writing and marketing techniques,” says Dr. Eric Sonstroem, the UOP English department chair. “I’m determined that my students really understand how content works on the Web, how it can be tested and measured, and how you can act on the data you get back.”
While classroom teaching is critical, UOP is also in the planning stage of a “hands on” content lab for its new campus in downtown San Francisco, opened last month near Twitter, Adobe, and in the city’s start-up SoMa district. It is envisioned that the lab will create, utilize and test software applications, conduct research, and educate students on content creation and analytic applications, and that students will intern within the tech community headquartered in the area.
UOP is certainly not the only institution of higher learning to address the data-driven marketplace needs, but the one I’m most familiar with, and other schools are today launching their own programs. Innovation is part of the U.S. DNA, and has been for centuries. Our institutions of higher education have been moving at a glacial pace but it’s beginning to thaw as the heat of market demand for digital sophisticates is “melting” the slow rate of change in higher education. We need to turn up the heat.
A recent article in VentureBeat said that marketing automation tools had only a 3 percent penetration rate at non-tech companies. Meanwhile, marketers are clamoring for ways to act upon data.
More or less, the weakest link in the chain of digital analytics has been the “make necessary changes” part. It’s now been several years since marketers began to understand that having the information alone really didn’t help the business. Recommendations became important. And after recommendations, then action.
Action is messy. It hasn’t had much to do, until recently, with automation. It required getting marketers, developers, creatives and business owners to agree on what changes were needed based on the data. And then the often too-laborious process of actually implementing the changes and trying to tell if there was a meaningful difference in the before and after states. Too often these efforts fell apart in partisan bickering between teams and refusal of many to take risks.
When we talk about marketing automation today, we are referring to SaaS offerings like Eloqua, Hubspot, Leadsius, Act-on and others that build a form of call-and-response matrix into marketing efforts. The easiest way to understand this is to compare it to what used to happen if you were reading a comic book when you were a kid, and saw an ad to “send away” for something either free or cheap. You would do that, and then you’d get more offers from the same company in the mail, as they hoped you’d soon spend more.
Much more dimensional and sophisticated versions of this are being played out by marketing automation tools, and according the the VentureBeat article, there’s plenty of room to grow.
A recent example of how one company is addressing a call for marketing automation is Tealium’s AudienceStream. Tealium already has a key foothold in the tag management industry, and that puts it at an important juncture of data collection. AudienceStream links the collected data from many sources (legacy of Tealium’s TMS) and allows the marketer to quickly set rules, thresholds and triggers that communicate via new APIs to marketing-action software already in the market. In other words, an AudienceStream powers an Eloqua. Once the rules are set, AudienceStream can communicate with a tool like Eloqua and help determine what message goes out to what user without continuing human intervention.
We’re not at the stage yet where entire site pages and app screens are being re-made on the spot based on very fresh data. We are at a stage where certain updatable modules on sites, and certain marketing messages can be automated and substituted based on data. The reason why this market sector has such growth potential is that it actually fixes a real problem.
While we’ve had lots of time to gnaw on old chestnuts like page views and unique visitors, we’ve hardly gotten to a point where we can say we’ve got organized, incremental methods that improve marketing velocity. And we know that most of the friction comes from friction between different teams with different agendas.
Marketing automation has no agenda except to respond to data and seek a return on marketing content. It frees up humans to do more strategic work. It may have only a small percentage of the market today, but as marketers get more and more familiar with successes based on these tools, that percentage is likely to begin growing rapidly in the near future.
Think of 2014 as the year when marketing automation finally got some of the recognition it deserves.