Marketing strategies that integrate Big Data, email, and cross-channel campaigns can reap big rewards. Among them are heightened brand awareness, upped conversion rates, and deeper customer satisfaction. Wrangling all those terabytes takes some work, but it’s work worth doing now more than ever, because what were once “new” communications, such as tweets, texts, and clickstreams, have become influential business resources.
The overall challenge for successful marketing campaigns is to pull it all together.
Direct Marketing News teamed up with sponsor email and cross-channel marketing solutions provider StrongMail in April to co-host a conversation among eight senior marketing professionals to explore Big Data and its challenges, opportunities, and paybacks. Among other issues, the roundtable examined such topics as detecting which email, cross-channel, and social marketing strategies drive business up or down; which lifetime data models do or don’t work for customer retention; transactional versus demographic data usage; and inhibitors to data access or analysis.
Editor-in-Chief Ginger Conlon, Direct Marketing News: Welcome to the StrongMail and Direct Marketing News roundtable on Big Data, multichannel, and email. I’ll begin with a big question:
What does Big Data mean in your company, what opportunities do you see for it in email and multichannel marketing?
Chad Ghastin (The New York Times): For our email programs it’s primarily two things: driving advertising revenue for our website and driving our subscription business. We look at behavioral data, and try to find relevant ways to target people with the right message to sign up for our emails. But let me just say…I think the term Big Data is a little bit of a misnomer—a bit consultant-speak, in some ways.
Ilana Rabinowitz (Lion Brand Yarn Company): We’re a small, family-owned business with a 135-year-old brand. I started our website and some email marketing 18 years ago, and now we’re heavily into social media and video marketing—we’ve just added some StrongMail modules for targeting people’s behavior. The website is about to be replatformed, but one of the nice things about it is that we know every move that everybody has made everywhere, everything they’ve ever done, every email they’ve sent, what they’ve said. We send a newsletter to something like 60, 65 million people a year, and we have crazy open rates, double and triple industry averages for the content. I really wanted to know from our Big Data, “What do we know?” When I started hearing some of what we know, I was amazed. Until you know what kind of information you have, you can’t even think about what could be in there.
Phil Thorn (Hiscox USA): We’re a business insurance company, and in terms of data we do a fair amount of acquisition, so we’re using data to deliver customized experiences online, and we’re starting to do a bit with email. But we’re not aggregating all that data and putting it into lifetime value models or really looking at the kind of retention piece and how we can use data to drive that. I’m interested in that particular piece of the puzzle.
Doug Jensen (Avon): We’re still a door-to-door, direct-selling company. Nineteen times a year—26 times a year in the United States—6.5 million representatives around the world place an order with us, and that order can contain data [of] up to 50 customers. The volume of data is quite big. Right now we have a pretty strong data mart. We’re looking at very top-level insights, in terms of what drives our business up and down and what the paybacks are for the things we do. But what we’re not really doing is using it for retention or churn analysis. We’re still [figuring] that out.
Like Chad, I also think Big Data’s a misnomer. I think it’s simply, we have data, and we have either predictive analytics or research.
Chris Loll (Wunderman): We’re part of WPP’s communications group, and focus heavily on data, direct, and digital. I would probably agree with the terminology the media and industry is adopting. There are so many forms of structured and unstructured data that are coming at us that it’s paralyzing people when they’re decision making. We own our own databases that [include] demographic, transactional, and behavioral data. We’re trying to understand how to match these things together to be more informed, to help us inform our strategies and our communications. And each of those things may be different by client.
Rafael Cardoso (Business & Legal Resources): I was brought on to launch a new initiative for…online media content. A large percentage of our business is still print subscribers to books, print newsletters. Over the past [few] years we’ve gone online with our subscription base, collecting a lot of data. Gathering all that data and using it to predict what our lifetime value would be online is where we’re having trouble. We silo decision making and the gathering of that data.
Michael Kildale (FreshDirect): My role is largely around marketing communications to customers—I’m in the integrated marketing group. We try to figure out [the] best ways to communicate to our customers. That’s what we’re really trying to tackle at FreshDirect. People come to us frequently, [but] we want to get people there more often. And when they are there, how do we get the best success figure? It’s a big challenge every one of us [here] has. There are a number of different ways to tackle it. We do offsite customer surveys to find out not just what customers purchase, but also to know why they’re not purchasing more, why they’re not purchasing at all, why they’re not purchasing a specific product.
Kara Trivunovic (StrongMail): I’m vice president of marketing services at StrongMail, responsible for the agency side of the business, the creative, production, strategy, and project management….
Regarding the conversation so far, it’s about finding the balance between how you want to take the data that you have and apply it to what you’re trying to accomplish, like using the behavioral data necessary to prove out your theory.
And I would argue that we’re not ready for Big Data. [Marketers] have a lot more information available to them today, and they feel like they have to use it. First you want to do the fundamentals right; make sure you have the right positioning, the right website, the right [customer] experience.
DMN: What opportunities do you see for integrating structured and unstructured data, what do you see as benefits and challenges, and what outcomes do you envision in your organization?
Thorn: We can combine research, analytics, and transactions data, and pool that to produce more tailored, personalized experiences. But for me, there are so many fundamentals that should come first: Have you got the right product? Have you got a compelling insight that’s really driving your customer communications? Is your website delivering a great user experience for everybody? Not just these micro segments that you might want to boil down to. All the basics should come first, before anybody even starts thinking about Big Data and all these very sexy, sophisticated algorithm and personalization platforms.
Rabinowitz: What’s fascinating to me about Big Data is that it’s actual [customer] behavior. I think that for most small and midsize companies Big Data is irrelevant. It’s like the Wild, Wild West of information. You can’t use it, you don’t have people who can analyze it, and if it’s there, you don’t know how to get to it. We [are] fortunate, because I have one person who’s got access to it who has a huge mind and who can really give us some answers from it. It’s unusual for a company our size.
Cardoso: If we had the data and an understanding of what customer segments we need to target and what that behavior looks like, we could invest more and we could really pull business from there. What we’re finding trouble with is that our close—from acquisition online to close—is anywhere from 90 to 120 days. In the meantime, there are a lot of untapped multichannel touchpoints: telesales, direct salespeople who are following up on email, events that we hold across the country.
Jensen: Instead of getting granular…we look holistically [for] everything that our data tells us. And, for me, multivariate regression does the best job of parsing through. If you have 100 variables going on, what things are driving sales up and down? When you stay at that high level, you’re basically saying, “Here’s what’s driving the business”—then you position yourself as a strategic partner.
So, we’re not into reporting. We collaborate with the different areas of marketing so we find things that they might not have asked for. We expect everyone in marketing to be able to report on their own data from their own system…. I think that’s something for people to think about when they’re structuring this.
Kildale: To me it’s about access. As an example, we used data last year to launch a same-day delivery service. We did it on a small scale. It was through research and anecdotal phone conversations from customer service and comments made through the email tool, response tools, and things like that where the same-day concept…kept coming up.
Trivunovic: Companies want to be able to look at what the behavioral data is saying to partner it with the information that they have about their customers already, the demographics. They’ve got background [information]. A lot of the challenge is in deciding what questions marketers want to answer first. The reality is that it doesn’t necessarily all have to culminate in one place at one moment in time. It informs different decisions.
DMN: How are you using Big Data to pull email and multichannel areas together to make decisions? Also talk about the biggest inhibitors to using more data.
Cardoso: We’re using data to build out segments and email, targeting keywords in our organic and paid search campaigns, telesales, and deployment setting to get transactional information and what I’d call back-end data to help to build campaigns. We don’t have a lot of actionable front-end data, other than opens and clicks and that kind of user interaction. I’d love to get some of that data that’s in between the front-end data and back-end conversion data to help us be more nimble and strategic with how to promote to our customers and make us more profitable.
Rabinowitz: We want to know what aspects of an email people are responding to, and based on that we could learn, say, that people like images of cats. It’s not just what links did they click on, but over time what kind of patterns do you see? Do people who read the blog open their emails at a higher rate than people who don’t read the blog? Do people who come to us from Facebook behave in a different way from people who engage with us on Twitter? And what does mean for our digital marketing.
Thorn: We’re getting more and more into the behavioral piece, what people do when they come through to the website. What search terms are they using? What pages are they visiting? What the conversion rate is like for this segment versus this segment. I’m enjoying it because we can actually start proving some ROI through A/B testing, multivariate testing, and through serving different messages to different segments. We’re showing senior management that we’ve run this test, done these five things, and this has worked. Here are the numbers. This is the ROI. And that makes conversations about data so much easier.
Kildale: We’ve done a much better job over the past couple of years, and not just within the marketing group…. The analytics team basically does both the reporting and analytics. The marketing group is responsible [for] developing a strategy and putting it out in the market. One can’t work without the other. The company would start failing because we’d just keep churning out the same [old] stuff.
Also, I think the barriers-to-purchase are something the data doesn’t give you. The question of whether somebody reading a blog being more likely to open an email is right. One of the things that we struggle with is information that leads to another question that you didn’t know to ask until you [get the data]. If you can get the “I believe” out of the conversation and insert the “I know,” it makes a big difference.
Trivunovic: A company has to make the cultural decision that it wants the organization to work together. A breakdown happens when the people responsible for bringing data elements together force the conversation with marketing ops. It’s not an easy task. You have to want and need the change to drive it.
I think part of the challenge is, once you have the data there and aligned you have to continue to drive that behavior within the organization…; to pull together as a team and bring that conversation to a much higher level. It’s not about what the social team is doing and what the email team is doing and what the Web team is doing. It’s very specific—to answer those business questions.
DMN: Creating a holistic customer view from multichannel data is no easy feat. What approaches are you taking that work effectively and what challenges are you having?
Loll: We see most of our clients thinking about the customer’s relationship with them only as a company. That’s a fairly narrow view of what we call the whole consumer. We believe that new sources—in particular things like clickstream data, mining unstructured data, and mapping data at an individual level—can drive incredible amounts of strategic value back to the organization. We’ve just completed an analysis for an iconic fashion brand and they’ve been building their segments and marketing based on attitudinal data that was informing their media choices and communications plan. Our analysis used their segments to reinterpret the marketplace: What are they actually consuming and how are they behaving? It naturally informs media channels, media planning, and buying. It also can inform communications, messaging, and creativity.
Thorn: I go down to my call center once a month and come away with a page of ideas and thoughts about how we can improve processes. We also get Web analytics data. You can find the issues and come up with solutions. Implementing those solutions is often tough: Have you got the technology bandwidth and the Web resources to go in and fix this stuff as quickly as you want? Data is great, but it’s what you do with this data that counts.
Ghastin: Our digital subscription business is taking off now, so we want to use data to inform our customer experience, those “moments of truth.” The question I want to answer is what touchpoints are most important to our subscribers? This informs us [of] how we need to realign our staffing and financial resources and deliver value efficiently and profitably. We have some pretty avid subscribers and our brand means a lot to them. And they actually do own our brand in many ways. It’s refreshing that you can use data to take some of the opinions out of the equation and inform the subscriber experience.
Trivunovic: The reality is that [all this multichannel data] is not a brand new thing. A lot of companies are going to have to change their stripes to make it work for them. It’s a different focus on how to leverage it, starting at the top, asking the big questions first, and letting the data help us answer those questions. Ultimately, the right decisions are about how we engage with the customer, not solely in a specific channel or down a specific path.
Kildale: The customer comes and just sees FreshDirect. That’s it. It’s a dangerous game if I start thinking about a display person versus a service person versus an email person. The customer doesn’t see it that way; they don’t care where they come in and how they came in.
DMN: Talk about some of your data goals; what are the obstacles to meeting them?
Cardoso: Resources are the first thing for us; building a foundation for our team to own [them]. Once we have a behavioral data set we can understand the customer online more than we do now, then use the data at a high level, more strategically, from print to online.
Thorn: I think we’re good at taking all these discreet data sources and understanding what customers do and what they need, and then segmenting them to drive acquisition. But we haven’t done anything yet about taking what we know about the customer and feeding that into lifetime models. I think the difficulty for us is just joining up all of that data…so we can really slice and dice our retention data by those variables to drive lifetime value—not just acquisition.
Ghastin: Some of the challenges I’ve been hearing about from colleagues outside my company are that companies think they can build this themselves. They believe that hiring five data scientists and somebody from NASA, and setting up shop in the back room, will solve their analytical obstacles. In my experience, you need specialized expertise from an outside company to help structure the data, and make you address things that you wouldn’t even consider because you’re so focused on building. There’s a lot of value in looking outside to the experts and not taking the “Let’s build this and see how it goes” approach.
Rabinowitz: That’s a good point. At [Lion Brand Yarn] we’re always trying to figure out how to do everything in house. Sometimes we hire a consultant to teach us and then we take it [from there]. But I think, especially for a company like ours, we would need some outside expertise.
Jensen: I’d like to underscore what I’m hearing about how Big Data can work. First, for me, is having an IT infrastructure where the data exists, which sounds like a lot of us are challenged with. [Second], I think having a top-down culture where from the CEO on down is saying, “We want to change how we use information, we want to have strategic decision making in our organization, and [we want] to focus on the customer. So all this is not data for data’s sake or information for information’s sake. It’s [about], how do we improve our relationship with the customer and get the customer what he or she needs? And that’s how we think through those three things.
